Abstract: An object detection method and an object detection apparatus are disclosed. In the object detection method, a bitstream of a progressively encoded JPEG image is fetched from an image supply device, wherein the bitstream includes a header and a subset of a plurality of discrete cosine transform (DCT) coefficients of the JPEG image, the subset of the DCT coefficients is decoded according to information in the header to reconstruct a coarse scan image of the JPEG image; and an object detection is performed on the coarse scan image to detect an object in the JPEG image.

Abstract: A sound source localization system includes a microphone array composed of a plurality of microphones each converting sound wave into a corresponding voice signal; a room shape estimator that determines a room shape including a location map and a corresponding template voice feature map composed of template voice features associated with a virtual sound source disposed at different locations respectively, and outputs a room reliability indicating confidence about the determined room shape; a lookup table (LUT) that pre-stores the location map and the corresponding template voice feature map; and a localizer that determines a location of a sound source according to the room reliability and similarity between a voice feature associated with the sound source and the template voice features of the template voice feature map.

Abstract: An anti-spoofing authentication system includes an image capture device that converts received light into a voice signal representing a captured image; a microphone that converts sound wave into a voice signal; a face detection device that detects a human face on the captured image, the face detection device including a lip detection device that detects lip features associated with the detected human face; a voice activity detection (VAD) device that detects presence of human voice in the voice signal; and an anti-spoofing device that detects a spoofing attack according to the detected lip features and the detected human voice.

Abstract: A sound source localization system includes a microphone array composed of a plurality of microphones each converting sound wave into a corresponding voice signal; a room shape estimator that determines a room shape including a location map and a corresponding template voice feature map composed of template voice features associated with a virtual sound source disposed at different locations respectively, and outputs a room reliability indicating confidence about the determined room shape; a lookup table (LUT) that pre-stores the location map and the corresponding template voice feature map; and a localizer that determines a location of a sound source according to the room reliability and similarity between a voice feature associated with the sound source and the template voice features of the template voice feature map.

Abstract: A method for detecting a specific object based on a specific model includes: capturing a set of images, wherein objects in each image include a desired object; transmitting the set of images to a cloud server; in response to found objects being obtained from the set of images based on at least one object detection algorithm in the cloud server, displaying the found objects fora user to confirm which object is desired; in response to the desired object being confirmed from the found objects that are displayed, loading the specific model of the desired object from the cloud server, wherein the specific model of the desired object is trained on the cloud server based on at least the set of images and related CNN algorithm; and performing the specific model to detect the specific object on a captured image.

Abstract: A voice activity detection (VAD) system includes a voice frame detector that detects a voice frame during which a voice signal is not silent; and a voice detector that detects presence of human speech according to the voice frame.

Abstract: A convolutional neural network (CNN) method includes determining a temporary buffer layer, which is located between a first layer and a final layer of a CNN system; performing convolutional operations from the first layer to the determined temporary buffer layer of the CNN system in a first stage to generate a feature map line according to partial input data of layers before the temporary buffer layer; and performing convolutional operations from the temporary buffer layer to the final layer of the CNN system in a second stage to generate a feature map.

Abstract: A method for detecting a specific object based on a specific model includes: capturing a set of images, wherein objects in each image include a desired object; transmitting the set of images to a cloud server; in response to found objects being obtained from the set of images based on at least one object detection algorithm in the cloud server, displaying the found objects fora user to confirm which object is desired; in response to the desired object being confirmed from the found objects that are displayed, loading the specific model of the desired object from the cloud server, wherein the specific model of the desired object is trained on the cloud server based on at least the set of images and related CNN algorithm; and performing the specific model to detect the specific object on a captured image.

Abstract: A monitor method for detecting an object includes capturing an image; calculating an initial probability of the object existing in the image; applying a mask to cover a first portion of the image if the initial probability is higher than a threshold; calculating a first probability of the object existing in the image excluding the first portion; and using at least the first probability to detect the location the object in the image.

Abstract: A monitoring device and an operation method thereof are provided to detect whether an object of interest appears in a video stream. The monitoring device includes a motion calculation circuit, a motion region determination circuit and a computing engine. The motion calculation circuit performs motion calculation on a current frame in the video stream to generate a motion map. The motion region determination circuit determines a motion region in the current frame according to the motion map. The motion region determination circuit notifies the computing engine with the motion region in the current frame. The computing engine performs an object of interest detection on the motion region in the current frame of the video stream to generate a detection result. The motion region determination circuit determines whether to ignore the motion region in a subsequent frame after the current frame according to the detection result.

Abstract: A monitor method for detecting an object includes capturing an image; calculating an initial probability of the object existing in the image; applying a mask to cover a first portion of the image if the initial probability is higher than a threshold; calculating a first probability of the object existing in the image excluding the first portion; and using at least the first probability to detect the location the object in the image.

Abstract: A monitoring device and an operation method thereof are provided to detect whether an object of interest appears in a video stream. The monitoring device includes a motion calculation circuit, a motion region determination circuit and a computing engine. The motion calculation circuit performs motion calculation on a current frame in the video stream to generate a motion map. The motion region determination circuit determines a motion region in the current frame according to the motion map. The motion region determination circuit notifies the computing engine with the motion region in the current frame. The computing engine performs an object of interest detection on the motion region in the current frame of the video stream to generate a detection result. The motion region determination circuit determines whether to ignore the motion region in a subsequent frame after the current frame according to the detection result.

Abstract: A convolutional neural network (CNN) method includes determining a temporary buffer layer, which is located between a first layer and a final layer of a CNN system; performing convolutional operations from the first layer to the determined temporary buffer layer of the CNN system in a first stage to generate a feature map line according to partial input data of layers before the temporary buffer layer; and performing convolutional operations from the temporary buffer layer to the final layer of the CNN system in a second stage to generate a feature map.