Abstract: A method and system for isolating the heart in a 3D volume, such as a cardiac CT volume, for patients with coronary artery bypasses is disclosed. An initial heart isolation mask is extracted from a 3D volume, such as a cardiac CT volume. The aortic root and ascending aorta are segmented in the 3D volume, resulting in an aorta mesh. The aorta mesh is expanded to include bypass coronary arteries. An expanded heart isolation mask is generated by combining the initial heart isolation mask with an expanded aorta mask defined by the expanded aorta mesh.

Abstract: In a method to determine an MR system-dependent phase information of a phase value in an MR phase image data set, initial variances are associated with image points of a first phase image data set, and for an additional phase image data set, a phase value and a variance of the phase value for the image points of the additional phase image data set are estimated by a linear transformation of the first phase image data set. The additional image data set is calculated by weighted addition of the phase values of the first phase image data set with the estimated phase values, with the weighting of the two phase values taking place in the addition dependent on the variance belonging to the respective phase value.

Abstract: An image processing apparatus for diagnostic imaging,which allows easy verification of a spatial position on an arbitrary cross section is provided.

Abstract: An Angiographic X-ray imaging system provides enhanced image feature visualization. At least one repository includes data representing a sequence of X-ray images of a portion of patient anatomy acquired over a time interval and signal data representing electrical activity of the heart of the patient over the time interval. An image data processor determines for individual images of the sequence of X-ray images, characteristics of a portion of the signal data associated with a corresponding image. The characteristics comprise at least one of, (a) a peak to peak value of a portion of the signal data associated with a corresponding image and (b) an average value of a portion of the signal data associated with the corresponding image. The image data processor selects a set of images exclusive of particular images from the sequence of X-ray images in response to the determined characteristics and generates an averaged image from the set of images.

Abstract: Methods and apparatuses process images. The method according to one embodiment accesses digital image data representing an image including an object; accesses reference data including a shape model relating to shape variation of objects from a baseline object, the objects and the baseline object being from a class of the object; and removes from the image an element not related to the object, by representing a shape of the object using the shape model.

Abstract: The present invention relates to a device and to a method for reducing in particular position-related gray value variations of image data sets of an object (18), in particular of a heart, which are recorded in various first layers (SA1, SA2, . . . , SAn) as respective first pictures (SP1, SP2, . . . , SPn) and in at least one second layer (LA1) intersecting with at least one first layer (SA1, SA2, . . . , SAn) as a respective second picture (LP1), comprising the following steps: registering the respective first pictures (SP1, SP2, . . . , SPn) of the first layers (SA1, SA2, . . . , SAn) relative to each other and with the at least one second picture (LP1) of the at least one second layer (LA1) in order to associate the first pictures (SP1, SP2, . . . , SPn) and the at least one second picture (LP1) of the object (18) with each other, and adjusting respective gray values of the object (18) in the at least one second picture to the respective gray values of the object (18) in the first pictures (SP1, SP2, .

Abstract: A rapid and efficient method and apparatus for detecting electrophysiologic, proarrhythmic, contractile, and other effects of substances such as compounds and drugs in native cellular cardiac preparations, the preparations representing an integrated cell-based pharmacologic response is disclosed. More specifically, a method to (1) rapidly and efficiently detect and verify the effects of chemicals, compounds and drugs on cardiac repolarization, contractility, and excitability using optically based techniques and customized simulation protocols, and (2) rapidly and efficiently screen and select compounds for electrophysiologic and proarrhythmic effects on cardiac myocytes is disclosed.

Abstract: An image processing program AP comprises: a step (S1) of reading time-lapse images; a step (S2) of extracting cells in a viewing image from each time point and performing labeling and cell association; and a step (S3) of determining the maturity of a cell included in the viewing image from a designated time point tc. In step (S4), an ID is assigned for a cell determined to be mature, and in step (S5), ID inheritance is performed by cell tracking of the time-lapse images in the backward or forward direction of time along the time axis. In step (S6), the number of unifications of cells is computed, and cells are classified on the basis of the inherited ID information, and the results of analysis are outputted in step (S7).

Abstract: A bank-note processing apparatus includes a debanding station and an image capturing device by which first and second information items applied to the band can be optically captured and digitally stored. The image capturing device includes at least one mirror, preferably two symmetrically arranged mirrors, in order for information items on the band that lie on different sides of the bank-note stack on the banded bank-note stack to be able to be optically captured simultaneously. The image capture can be effected before or after the removal of the band from the bank-note stack.

Abstract: A jig 20 that is installed on a device for measuring dimensions of a workpiece 10 based on images obtained by taking pictures of the workpiece 10 with a camera K includes a chuck mechanism 50 holding the workpiece K, a first rotation drive mechanism 80 that rotates the chuck mechanism 50 around a predetermined first rotation axis A, a first base 30 holding the first rotation drive mechanism 80, a second rotation drive mechanism 40 that rotates the first base 30 around a second rotation axis B orthogonal to the first rotation axis A, and a second base 21 holding the second rotation drive mechanism 40. The first base 30 has an opening 31A in the area around the first rotation axis A such that the workpiece 10 held by the chuck mechanism 50 is seen from the back of the first base 30 through the opening 31A.

Abstract: Disclosed herein is an image processing apparatus including: a depth-information extraction section; a luminance extraction section; a contrast extraction section; a gain generation section; and a correlation estimation section.

Abstract: An image such as a depth image of a scene may be received, observed, or captured by a device and a model of a user in the depth image may be generated. The background of a received depth image may be removed to isolate a human target in the received depth image. A model may then be adjusted to fit with in the isolated human target in the received depth image. To adjust the model, a joint or a bone may be magnetized to the closest pixel of the isolated human target. The joint or the bone may then be refined such that the joint or the bone may be further adjusted to a pixels equidistant between two edges the body part of the isolated human target where the joint or bone may have been magnetized.

Abstract: Presented is a method of automatically performing an action, based on graphical input. The method comprises: receiving, for a user, an input image; comparing the input image with the contents of a user-customized database comprising a plurality of records, each record representing a predefined class of image, wherein the user has previously associated records in the database with respective specified actions; attempting to recognize the image, based on the similarity of the input image to one of the predefined classes of image represented in the user-customised database; and if the image is recognized, performing the action previously associated by the user with the class. Also presented is apparatus for recognizing an image and a method of constructing a user-customized database.

Abstract: An information processing apparatus includes: a distinguishing unit which, by using an ensemble classifier, which includes a plurality of weak classifiers outputting weak hypotheses which indicates whether a predetermined subject is shown in an image in response to inputs of a plurality of features extracted from the image, and a plurality of features extracted from an input image, sequentially integrates the weak hypotheses output by the weak classifiers in regard to the plurality of features and distinguishes whether the predetermined subject is shown in the input image based on the integrated value. The weak classifier classifies each of the plurality of features to one of three or more sub-divisions based on threshold values, calculates sum divisions of the sub-divisions of the plurality of features as whole divisions into which the plurality of features is classified, and outputs, as the weak hypothesis, a reliability degree of the whole divisions.

Abstract: A system for detecting a global harmful video includes: a video determination policy generation unit for determining harmfulness of learning video segments from video learning information to analyze occurrence information of harmful learning video segments, and generating a global harmfulness determination policy based on the occurrence information; and a video determination policy execution unit for determining harmfulness of input video segments from information of an input video to analyze occurrence information of harmful input video segments, and determining whether the input video is harmful or not based on the occurrence information of the harmful input video segments and the global harmfulness determination policy.

Abstract: In an image processing device and method, program, and recording medium of the present invention, high frequency components of a low quality image and a high quality image included in a studying image set are extracted, and an eigenprojection matrix and a projection core tensor of the high frequency components are generated in a studying step. In a restoration step, a first sub-core tensor and a second sub-core tensor are generated based on the eigenprojection matrix and the projection core tensor of the high frequency components, and a tensor projection process is applied to the high frequency components of an input image to generate a high quality image of the high frequency components. The high quality image of the high frequency components is added to an enlarged image obtained by enlarging the input image to the same size as an output image.

Abstract: Unevenness in density (line noise) resulting from color misregistration among RGB components in a halftone region or a line screen region is suppressed. With respect to a document that is discriminated as a printed photo document or a text/printed photo document that includes black halftones by a document type discrimination section, or to a region segmented into a black halftone region by a segmentation section, a non-uniform image suppressing operation mode is executed. In the non-uniform image suppressing operation mode, a black generation and under color removal section does not carry out a black generation and under color removal process on a pixel in which a minimum value of the image data of the plurality of color components is at least equal to or less than a predetermined threshold.

Abstract: The compression processing unit 30 of the image processing apparatus makes the color conversion process section 50 to the layout analysis processing section 53 extract the text string region, and makes the foreground layer generating section 54 extract the foreground pixel from the text string region to generate the foreground layer. The line detection process section 52 extracts the line art from the input image data, and removes the pixel block configuring the line art from the foreground pixel-block mask generated by the foreground extracting section 51. The layout analysis processing section 53 estimates the direction of the text string in accordance with the distance of the foreground regions and the number of connected foreground regions, and merges the foreground regions connected in the estimated direction, to extract the text string region.

Abstract: A system and method for identifying an unknown substance in a sample comprising multiple entities. A method may comprise generating a RGB image representative of a sample and assessing said RGB image to identify at least one region of interest. This region of interest may he assessed to generate a spatially accurate wavelength resolved image, which may be a hyperspectral image. This spatially accurate wavelength resolved image may comprise a fluorescence, Raman, near infrared, short wave infrared, mid wave infrared and/or long wave infrared image. This spatially accurate wavelength resolved image may be assessed to identify said unknown substance. A system may comprise: a reference database, a first detector for generating an RGB image, a second detector for generating a spatially accurate wavelength resolved image, and a means for assessing said RGB image and said spatially accurate wavelength resolved image.

Abstract: An automated and extensible system is provided for the analysis and retrieval of images based on region-of-interest (ROI) analysis of one or more true objects depicted by an image. The system uses an ROI database that is a relational or analytical database containing searchable vectors representing images stored in a repository. Entries in the ROI database are created by an image locator and ROI classifier that locate images within the repository and extract relevant information to be stored in the ROI database. The ROI classifier analyzes objects in an image to arrive at actual features of the true object. Graphical searches may also be performed.

Abstract: In an image processing apparatus, a white balance calculating unit is configured to calculate a first coordinate value that indicates a white balance of an image in a color space. A determining unit is configured to determine whether a first distance between the first coordinate value and a reference coordinate value is greater than a prescribed threshold distance. An adjustment amount setting unit is configured to set the adjustment amount to a first adjustment amount when the determining unit determines that the first distance is smaller than or equal to the prescribed threshold distance, the first adjustment amount corresponding to the first distance, the adjustment amount setting unit setting the adjustment amount to a second adjustment amount when the determining unit determines that the first distance is greater than the prescribed threshold distance, the second adjustment amount corresponding to a second distance smaller than the first distance.

Abstract: A first generation unit generates a RAW image reduced in the horizontal direction from a RAW image including pixels each having single-color signal information, by limiting the band of spatial frequencies in the horizontal direction and decreasing the number of pixels in the horizontal direction. The RAW image reduced in the horizontal direction is stored in a line memory. A second generation unit generates a reduced RAW image by decreasing, at least in the vertical direction, the number of pixels of the RAW image reduced in the horizontal direction. This makes it possible to generate a reduced RAW image using little memory capacity.

Abstract: A device for recognizing and locating objects in an image by scanning detection windows comprises a data stream architecture designed in pipeline form for concurrent hardware tasks and includes means for generating a descriptor for each detection window, a histogram determination unit determining a histogram of orientation gradients for each descriptor, and N processing units in parallel, capable of analyzing the histograms as a function of parameters associated with the descriptors to provide a partial score representing the probability that the descriptor concerned contains at least part of the object to be recognized, the sum of the partial scores of each detection window providing a global score representing the probability that the detection window contains the object to be recognized.

Abstract: Geometric method of transforming a first two-dimensional image into a second two-dimensional image through an image processing applied to the first images or to the second image. In this method, one of said first and second images is divided into several surfaces, each of the surfaces of the divided image is transformed by a bilinear transformation specific to each surface, and the transformed surfaces are grouped together.

Abstract: A “Text Rectifier” provides various techniques for processing selected regions of an image containing text or characters by treating those images as matrices of low-rank textures and using a rank minimization technique that recovers and removes image deformations (e.g., affine and projective transforms as well as general classes of nonlinear transforms) while rectifying the text or characters in the image region. Once distortions have been removed and the text or characters rectified, the resulting text is made available for a variety of uses or further processing such as optical character recognition (OCR). In various embodiments, binarization and/or inversion techniques are applied to the selected image regions during the rank minimization process to both improve text rectification and to present the resulting images of text to an OCR engine in a form that enhances the accuracy of the OCR results.

Abstract: An image of a known text sample having a text type is generated. The image of the known text sample is input into each OCR engine of a number of OCR engines. Output text corresponding to the image of the known text sample is received from each OCR engine. For each OCR engine, the output text received from the OCR engine is compared with the known text sample, to determine a confidence value of the OCR engine for the text type of the known text sample.

Abstract: A server system receives a visual query from a client system distinct from the server system. The server system performs optical character recognition (OCR) on the visual query to produce text recognition data representing textual characters, including a plurality of textual characters in a contiguous region of the visual query. The server system scores each textual character in the plurality of textual characters in accordance with the geographic location of the client system. The server system identifies, in accordance with the scoring, one or more high quality textual strings, each comprising a plurality of high quality textual characters from among the plurality of textual characters in the contiguous region of the visual query. Then the server system retrieves a canonical document having the one or more high quality textual strings and sends at least a portion of the canonical document to the client system.

Abstract: An image processing apparatus includes a cutout position extraction unit, a character candidate extraction unit, a graph generation unit, a link value generation unit, a path selection unit and an output unit. The cutout position extraction unit extracts a cutout position. The character candidate extraction unit recognizes each character for each character image divided by the cutout position and extracts a plurality of character candidates for each recognized character. The graph generation unit sets each of the plurality of extracted character candidates as a node and generates a graph by establishing links between the nodes of adjacent character images. The link value generation unit generates a link value based on a value of character-string-hood representing a relationship between character candidates. The path selection unit selects a path in the generated graph based on the link value. The output unit outputs a character candidate string in the selected path.

Abstract: An image registration system for aligning first and second images. The novel system includes a first system for extracting a region of interest (ROI) from each image and a second system for coarsely aligning the regions of interest. The first system determines the size and location of the ROI based on the number of features contained within the region. The size of the ROI is enlarged until a number of features contained in the ROI is larger than a predetermined lower bound or until the size is greater than a predetermined upper bound. The second system computes a cross-correlation on the regions of interest using a plurality of transforms to find a coarse alignment transform having a highest correlation. The image registration system may also include a third system for performing sub-pixel alignment on the regions of interest.

Abstract: The present invention may provide a method for image-based identification. The method may include providing a digital photo of an unidentified item; transmitting, over a network, the digital photo to an identification service; in response to transmitting the digital photo, receiving, over the network, item information from the identification service, wherein the item information includes textual identification information about the item; and displaying the textual identification information.

Abstract: A method for determining with a physical image processing device a visual balance of an image includes assigning a visual weight point to each of a plurality of visual elements within the image with the image processing device, each visual weight point having a weight value based on visual properties associated with the visual element, and determining the visual balance of the image with the image processing device by measuring a vector value at a center of the composition, the vector value being based on a distance of each visual weight point from the center and the weight value associated with each visual weight point.

Abstract: At least one region of interest within an image is determined, step 201. The image is resized to a target image size step 207. If at least one region of interest is determined the image is cropped to each of the at least one region of interest, step 205 and the cropped at least one region of interest is resized to the target image size. The re-sized image and the resized cropped at least one region of interest are stored for display, step 209.

Abstract: A picture processing device (100) includes: a mask generation unit (110x) which generates a masking picture (110aI) based on an edge in a first picture (101a); a feature point masking unit (111x) which masks extracted feature points with the generated masking picture (110aI); a rotation matrix calculation unit (112) which calculates a positional change between the first picture and a second picture (101a, 101b) using a feature point selected from the feature points as a result of the masking; and a picture correction unit (113) which uses the calculated change.

Abstract: A dense guide image or signal is used to inform the reconstruction of a target image from a sparse set of target points. The guide image and the set of target points are assumed to be derived from a same real world subject or scene. Potential discontinuities (e.g., tears, edges, gaps, etc.) are first detected in the guide image. The potential discontinuities may be borders of Voronoi regions, perhaps computed using a distance in data space (e.g., color space). The discontinuities and sparse set of points are used to reconstruct the target image. Specifically, pixels of the target image may be interpolated smoothly between neighboring target points, but where neighboring target points are separated by a discontinuity, the interpolation may jump abruptly (e.g., by adjusting or influencing relaxation) at the discontinuity. The target points may be used to select only a subset of the discontinuities to be used during reconstruction.

Abstract: The method includes calculating similarities between a plurality of pixel signals of a depth pixel and a plurality of pixel signals of neighbor depth pixels neighboring the depth pixel, calculating a weight of each of the neighbor depth pixels using the similarities, calculating a weight of the depth pixel using the weights of the respective neighbor depth pixels, and determining a denoised pixel signal using the weights of the respective neighbor depth pixels and the weight of the depth pixel.

Abstract: The system according to the present invention includes a server including a map data storage device, which stores digital map data, a plurality of tile map data generated from the digital map data, and solid-compressed map data compressed by solid compression from the plurality of tile map data, a solid-compressed map data generation device, which generates the plurality of tile map data from the digital map data stored in the map data storage device, generates the solid-compressed map data by compressing the plurality of tile map data by solid compression, and stores the generated solid-compressed map data in the map data storage device, and a solid-compressed map data providing device, which transmits the solid-compressed map data to a client through a network in response to a request from the client. Therefore, it is possible to reduce the frequency of network access and the amount of data transmitted.

Abstract: A quality level determining the extent to which each image file is compressed is automatically computed for each image file in a set to ensure that the total size of the compressed image files does not exceed a predefined limit. The compressed size of each image file is initially determined when compressed at a predefined minimum acceptable level and at a nominal level. The relative complexity of the image files is determined based upon their high frequency energy content. As a function of the image file complexity, and starting with the compressed sizes initially determined, the appropriate quality level is determined for compressing each of the image files in an iterative process that ensures the total size of the compressed image files does not exceed the predefined limit, while retaining acceptable quality. Thus, a set of image files can be compressed optimally to fit within a limited storage.

Abstract: A high-frequency integrator acquires a first integrated value by integrating a high-frequency component of first image data. A corrector handles second image data, which is obtained when an image processor subjects the first image data to an image process affecting frequency characteristics, and acquires a second integrated value by correcting the first integrated value acquired by the high-frequency integrator in accordance with a change in the frequency characteristics that is brought about by the image process. An encoder calculates, in accordance with the second integrated value acquired by the corrector, a quantization scale for compressing the second image data acquired by the image processor to a predefined number of bytes at once, and compresses the second image data accordingly.

Abstract: An image conversion device is disclosed, having an image converting circuit for receiving a first and a second image frames of a first format and generating a third and a fourth image frames of a second format; and a signal generating circuit coupled with the image converting circuit for generating a plurality of first synchronization signals having a substantially fixed period and one or more second synchronization signals for the third and the fourth image frames, wherein each of the second synchronization signals is synchronized with one of the first synchronization signals and the third image frame contains at least one more first synchronization signal than the fourth image frame.

Abstract: Methods for scaling an image taken an optimal exposure time to a selected exposure time, generally comprising, determining a dark pixel intensity of an imaging device; acquiring a first image at an optimal exposure time; and adjusting a pixel intensity of one or more pixels in the first image, based at least in part on the dark pixel intensity, for a second exposure time that is different from the optimal exposure time.

Abstract: The fries and condiment holder is a combination of two open containers. The larger of the two open containers provides the user with a place to store a food item found at fast food restaurants such as French fries. The smaller of the two open containers is attached to the larger container by a nontoxic adhesive because the present invention makes contact with edible food. The smaller container provides the user with a place to store a condiment. The smaller container is also collapsible and can be flushed into larger container. The user obtains the present invention with the food item in the larger container and with the small container collapse. The user can choose to expand and fill the smaller container with a condiment if the user wants to dip his/her food item in the condiment.

Abstract: The present invention relates to a material suitable for a ply of a sack, comprising a porous sheet provided with a coating on at least one of its surfaces, wherein said coating comprises at least one polyolefin. Further, the present invention relates to a sack comprising a ply, which comprises such a material. The present invention also relates to a method of forming a coating on a porous sheet, comprising the steps of: providing a porous sheet and a dispersion comprising at least one polyolefin; applying said dispersion on at least one surface of said sheet; and optionally, heating said sheet to a temperature above the melting temperature of said at least one polyolefin. Moreover, the present invention relates to methods for manufacturing single and two ply sacks and the use of at least one polyolefin as a water vapour barrier coating on a material suitable for a ply of a sack.

Abstract: Methods for creating multi-layered bags with increased or maintained strength involve forming a ribbed pattern in one or more film layers of a multi-layered bag. The method also includes non-continuously laminating the film layers of the multi-layered bag together. In one or more implementation, a transverse direction ring rolling process forms the ribbed pattern and bonds the film layers together. In one or more additional implementations, the ribbed pattern and lamination are formed separately. Still further implementations include forming network patterns one or more film layers of a multi-layered bag. Multi-layered bags formed in accordance with one or more implementations of the present invention include one or more of increased strength or reduced basis weight.

Abstract: The present invention was directed to an expandable package which facilitates packaging of compressible articles, such as sanitary products, disposable diapers and the like. The expandable package includes an outer bag body, and at least one inner expansion element, which is maintained in a folded condition so that articles within the package are maintained in a compressed state. The package includes at least one preferentially weakened area in the outer bag body, which when opened, permits the inner expansion elements to unfold, thus facilitating convenient separation and removal of individual articles from within the package by consumers.

Abstract: A rolling element retainer chain provides a row of rolling elements smooth running in a recirculation passage and keeps a predetermined spacing distance between two adjacent rolling elements and the rolling elements be separated without contact to each other; the rolling element retainer chain comprises at least one thin flexible metal strip having a row of longitudinal arranged holes; a row of cross arranged latten between holes are formed thereof; and a number of spacers, incorporated on the cross arranged latten; the strength of the flexible metal strip is much higher than that of the resin materials conventionally used and can endure higher tension without broken, and thus a longer operating duration can therefore be achieved; both ends on the flexible metal strip can be overlapped wholly fixed without interference because of thinner thickness thereof, and a close loop rolling element retainer chain is formed thereof.

Abstract: Disclosed is a coupling structure of a bearing for an EPS motor to eliminate a gap caused by a tolerance between the bearing and a pocket of a housing of the EPS motor, thereby improving assembling precision and reducing noise, the coupling structure supporting rotation of a rotating shaft through the bearing coupled to a lower portion of the housing and including a bearing pocket bent upwards from a lower surface of the housing and having a cylindrical space to support an outer circumference of the bearing at an inner wall thereof, whereby a gap between the inner wall of the bearing pocket and the outer circumference of the bearing is eliminated by deforming the inner wall towards an inner circumference thereof by caulking, coupling force between the bearing and the housing can be increased, and thereby reliability in assembly and operation of the motor is improved.

Abstract: The invention concerns a drum motor comprising a drum tube, a spatially fixed shaft in the interior of the drum tube. To provide a drum motor in particular for use in hygiene areas a drum motor (10) according to the invention comprises: a drum tube (11), a spatially fixed shaft (13) in the interior of the drum tube (11), an inner cover (13) for accommodating a bearing (14) to support the drum tube (11) rotatably about the shaft (13) and for internally sealing off the interior (12), and an outer cover (30) for covering a drum tube end and for externally sealing off the interior (12).

Abstract: An arm bearing for an articulated arm awning comprises a bearing block which is fastened to a supporting block so as to be pivotable about a tilting axis and which, at one end, has at least one bearing eye for receiving a bearing axis of an awning arm in a manner rotatable about an axis of rotation, furthermore a tilting angle limiting element for limiting a tilting angle between the supporting block and the bearing block, wherein a first end of the tilting angle limiting element is connected to the supporting block, and a second end of the tilting angle limiting element is designed as a stop which interacts with a counterstop on the bearing block, and a locking slide which fixes the second end of the tilting angle limiting element and the counterstop on each other, wherein the locking slide is movable in a translatory manner by means of a rotational movement of the awning arm about the axis of rotation.

Abstract: An electro-mechanical fuse integrated into the enclosure of a bearing assembly of a generator or like component. Under normal operation, the electro-mechanical fuse provides continuity through a bearing sensing circuit. However, ball cage damage causes the fuse to open and, in turn, the bearing sensing circuit indicates bearing wear. Thus, damage to other generator components is greatly reduced because the generator is removed for maintenance in a timely manner.

Abstract: A radial shaft seal includes a seal ring connected to a support ring and having a seal lip that extends radially inwardly. A tension spring urges the seal lip radially inward against a seal raceway during operation of the radial shaft seal. A retaining device retains the tension spring and is radially supported on a rigid inner circumferential wall that radially surrounds the tension spring. The retaining device may be configured to at least one of: (i) limit a widening of the diameter of tension spring caused by centrifugal force to a predetermined upper limit and (ii) prevent the tension spring from flying off during operation at high rotational speeds.