Abstract: Embodiments of the present disclosure disclose a method and apparatus for processing an image. A specific embodiment of the method includes: acquiring a feature map of a target image, where the target image contains a target object; determining a local feature map of a target size in the feature map; combining features of different channels in the local feature map to obtain a local texture feature map; and obtaining location information of the target object based on the local texture feature map.

Abstract: A method and apparatus for detecting text regions in an image, a device, and a medium are provided. The method may include: detecting, based on feature representation of an image, a first text region in the image, where the first text region covers a text in the image, a region occupied by the text being of a certain shape; determining, based on a feature block of the first text region, text geometry information associated with the text, where the text geometry information includes a text centerline of the text and distance information of the centerline from the upper and lower borders of the text; and adjusting, based on the text geometry information associated with the text, the first text region to a second text region, where the second text region also covers the text and is smaller than the first text region.

Abstract: Embodiments of the present disclosure provide a method and apparatus for generating an image. The method may include: receiving a first image including a face input by a user in an interactive scene; presenting the first image to the user; inputting the first image into a pre-trained generative adversarial network in a backend to obtain a second image output by the generative adversarial network; where the generative adversarial network uses face attribute information generated based on the input image as a constraint; and presenting the second image to the user in response to obtaining the second image output by the generative adversarial network in the backend.

Abstract: A method for generating a sample image includes: obtaining an initial image size of an initial image; obtaining a plurality of reference images by processing the initial image based on different reference processing modes; obtaining an image to be processed by fusing the plurality of reference images; and determining a target sample image from images to be processed based on the initial image size.

Abstract: The present application discloses a video-based human behavior recognition method, apparatus, device and storage medium, and relates to the technical field of human recognitions. The specific implementation scheme lies in: acquiring a human rectangle of each video frame of the video to be recognized, where each human rectangle includes a plurality of human key points, and each of the human key points has a key point feature; constructing a feature matrix according to the human rectangle of the each video frame; convolving the feature matrix with respect to a video frame quantity dimension to obtain a first convolution result and convolving the feature matrix with respect to a key point quantity dimension to obtain a second convolution result; inputting the first convolution result and the second convolution result into a preset classification model to obtain a human behavior category of the video to be recognized.

Abstract: A method, an electronic device and a storage medium are provided. The method may include: acquiring a to-be-inspected image; inputting the to-be-inspected image into a pre-established vehicle detection model to obtain a vehicle detection result, where the vehicle detection result includes category information, coordinate information, coordinate reliabilities, and coordinate error information of detection boxes, and the vehicle detection model is configured for characterizing a corresponding relationship between images and vehicle detection results; selecting, based on the coordinate reliabilities of the detection boxes, a detection box from the vehicle detection result for use as a to-be-processed detection box; and generating, based on coordinate information and coordinate error information of the to-be-processed detection box, coordinate information of a processed detection box.

Abstract: A method and device for detecting a vehicle pose, relating to the fields of computer vision and automatic driving. The specific implementation solution comprises: inputting a vehicle left view point image and a vehicle right view point image into a part prediction and mask segmentation network model, and determining foreground pixel points and part coordinates thereof in a reference image; converting coordinates of the foreground pixels in the reference image into coordinates of the foreground pixels in a camera coordinate system so as to obtain a pseudo-point cloud, and fusing part coordinate of the foreground pixels and the pseudo-point cloud to obtain fused pseudo-point cloud; and inputting the fused pseudo-point cloud into a pre-trained pose prediction model to obtain a pose information of the vehicle to be detected.

Abstract: Embodiments of the present disclosure provide a video processing method, a video processing device and a related non-transitory computer readable storage medium. The method includes the following. Frame sequence data of a low-resolution video to be converted is obtained. Pixel tensors of each frame in the frame sequence data are inputted into a pre-trained neural network model to obtain high-resolution video frame sequence data corresponding to the video to be converted output by the neural network model. The neural network model obtains the high-resolution video frame sequence data based on high-order pixel information of each frame in the frame sequence data.

Abstract: A method and an apparatus for generating vehicle damage information are provided. The method includes: acquiring a to-be-processed vehicle image; for a target detection model in at least one pre-trained target detection model: inputting the vehicle image to the target detection model to generate a suspected damage area detection result; and determining a location of a suspected damage area in the vehicle image based on the generated suspected damage area detection result. A mechanism for detecting a suspected damage area is provided based on the target detection model, improving the vehicle damage assessment efficiency.

Abstract: The present disclosure provides a pedestrian re-identification method and apparatus, computer device and readable medium. The method comprises: collecting a target image and a to-be-identified image including a pedestrian image; obtaining a feature expression of the target image and a feature expression of the to-be-identified image respectively, based on a pre-trained feature extraction model; wherein the feature extraction model is obtained by training based on a self-attention feature of a base image as well as a co-attention feature of the base image relative to a reference image; identifying whether a pedestrian in the to-be-identified image is the same pedestrian as that in the target image according to the feature expression of the target image and the feature expression of the to-be-identified image.

Abstract: Embodiments of the present disclosure disclose a method and apparatus for recognizing a body movement. A specific embodiment of the method includes: sampling an input to-be-recognized video to obtain a sampled image frame sequence of the to-be-recognized video; performing key point detection on the sampled image frame sequence by using a trained body key point detection model, to obtain a body key point position heat map of each sampled image frame in the sampled image frame sequence, the body key point position heat map being used to represent a probability feature of a position of a preset body key point; and inputting body key point position heat maps of the sampled image frame sequence into a trained movement classification model to perform classification, to obtain a body movement recognition result corresponding to the to-be-recognized video.

Abstract: A method for video frame interpolation, a related electronic device and a storage medium is disclosed. A video is obtained. An (i?1)th frame and an ith frame of the video are obtained. Visual semantic feature maps and depth maps of the (i?1)th frame and the ith frame are obtained. Frame interpolation information is obtained based on the visual semantic feature maps and the depth maps. An interpolated frame between the (i?1)th frame and the ith frame is generated based on the frame interpolation information and the (i?1)th frame and is inserted between the (i?1)th frame and the ith frame.

Abstract: The present disclosure provides a human body recognition method and apparatus, and a storage medium, the method comprising: determining a coordinate of a target person in a three-dimensional space according to images containing the target person collected by at least two cameras; calculating back-projection errors of the target person under different cameras respectively according to the coordinate of the target person in the three-dimensional space; determining whether the cameras have a human body recognition error according to the back-projection errors of the cameras; when a camera has the human body recognition error, performing re-recognition of the target person under the camera by using person re-identification ReID, until the back-projection errors of all the cameras containing the target person are not greater than a preset threshold. The present disclosure can improve accuracy of the human body recognition result effectively.

Abstract: A human body tracking method, apparatus, and device, and a storage medium. The method includes: obtaining a current frame image captured by a target photographing device at a current moment; detecting each human body in the current frame image to obtain first position information of the each human body in the current frame image; calculating second position information of a first human body in the current frame image; determining target position information of the each human body in the current frame image according to the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and pedestrian features of all tracked pedestrians stored in a preset list.

Abstract: Provided are a training method and apparatus for a human keypoint positioning model, a human keypoint positioning method and apparatus, a device, a medium and a program product. The training method includes determining an initial positioned point of each of keypoints; acquiring N candidate points of each keypoint according to a position of the initial positioned point; extracting a first feature image, and forming N sets of graph structure feature data according to the first feature image and the N candidate points; performing graph convolution on the N sets of graph structure feature data to obtain N sets of offsets; correcting initial positioned points of all the keypoints to obtain N sets of current positioning results; and calculating each set of loss values according to labeled true values of all the keypoints and each set of current positioning results, and performing supervised training on the positioning model.

Abstract: A method and an apparatus for human behavior recognition, and a storage medium, the method includes: obtaining a human behavior video captured by a camera; extracting a start point and an end point of a human motion from the human behavior video, where the human motion between the start point and the end point corresponds to a sliding window; determining whether the sliding window is a motion section; and if the sliding window is a motion section, anticipating a motion category of the motion section using a pre-trained motion classifying model. Thus, accurate anticipation of a motion in a human behavior video captured by a camera is realized without human intervention.

Abstract: A method, apparatus, device, and storage medium for predicting the number of people of a dense crowd, including: converting a first image, in which the number of people is to be determined, into a corresponding first thermodynamic chart according to a thermodynamic chart conversion model; and determining the number of people in the first image according to the first thermodynamic chart, wherein the thermodynamic chart conversion model is obtained by training according to a pre-marked second image and a thermodynamic chart corresponding to each second image, thereby achieving prediction of the number of people of a dense crowd, improving the accuracy in predicting the number of people of the dense crowd while improving management efficiency.

Abstract: Embodiments of the present disclosure provide an object tracking method and an apparatus. The method includes: obtaining multiple frames of first images shot by a first camera apparatus and a first shooting moment of each frame of the first images, where the first images include a first object; obtaining multiple frames of second images shot by a second camera apparatus and a second shooting moment of each frame of the second images, where the second images include a second object; obtaining a distance between the first camera apparatus and the second camera apparatus; and judging whether the first object and the second object are the same object according to the multiple frames of the first images, the first shooting moment of each frame of the first images, the multiple frames of the second images, the second shooting moment of each frame of the second images and the distance.

Abstract: A method, apparatus and a system for human body tracking processing, where an apparatus for video collection processing in the system has a built-in intelligent chip, and before uploading video data to a cloud server, the intelligent chip performs a pre-processing on the video data, retains a key image frame and performs a human body detection and a tracking processing on the key image frame by using human body detection tracking algorithm to acquire a first human body detection tracking result. Afterwards, the intelligent chip sends the first human body detection tracking result to the cloud server, so that the cloud server performs a human body re-identification algorithm processing and/or three-dimensional reconstruction algorithm processing on the first human body detection tracking result to acquire a second human body detection tracking result.

Abstract: A method and an apparatus for classifying a video are provided. The method may include: acquiring a to-be-classified video; extracting a set of multimodal features of the to-be-classified video; inputting the set of multimodal features into a post-fusion model corresponding to each modal respectively, to obtain multimodal category information of the to-be-classified video; and fusing the multimodal category information of the to-be-classified video, to obtain category information of the to-be-classified video. This embodiment improves the accuracy of video classification.