Techniques for processing images of documents captured using a mobile device are provided. The images can include different sides of a document from a mobile device for an authenticated transaction. In an example implementation, a method includes inspecting the images to detect a feature associated with a first side of the document. In response to determining an image is the first side of the document, a type of content is selected to be analyze on the image of the first side and one or more of regions of interests (ROIs) are identified on the image of the first side that are known to include the selected type of content. A process can include receiving a sub-image of the image of the first side from the preprocessing unit, and performing content detection test on the sub-image.

Identification of ships from camera images is performed by computing sets of reference silhouette images for respective reference ships for use to match the reference silhouette images with a camera image of an unidentified ship. A computer forms the set of the reference silhouette images for a reference ship is performed automatically by a processing system on the basis of at least one silhouette image of the respective reference ship. The computer generates a three dimensional model of the reference ship that represents distances of an outer surface of the reference ship to a plane on which the captured silhouette image has been mapped, the three dimensional model representing the distances as a function of position in said plane. A deck line of the reference ship is detected based on the camera image. The represented distances to the outer surface are set and/or adapted dependent on position in said plane relative to the detected deck line. The represented distances for positions in said plane below the detected deck line are set and/or adapted convergently toward a bow of the ship in the silhouette image. The reference silhouette images are generated by computing perspective projections of the three dimensional model onto image planes for respective viewpoints in a three dimensional array of viewpoints in ranges of distance relative to the reference ship, horizontal direction from the reference ship to the viewpoint and height relative to the reference ship and in the three dimensional model.

Provided are methods, apparatus, and computer-readable mediums for tracking objects that intersect with an exclusion zone defined for a scene being captured by a video camera. An exclusion zone can delineate an area of a video frame where background objects may be moving. The exclusion zone informs an object tracking system that objects within the exclusion zone should not be tracked. In various implementations, the object tracking system can determine that a bounding box for a blob intersects with the exclusion zone. The object tracking system can further, based on the bounding box intersecting with the exclusion zone, prevent outputting of a blob tracker associated with the blob.

A method for dynamically quantizing feature maps of a received image. The method includes convolving an image based on a predicted maximum value, a predicted minimum value, trained kernel weights and the image data. The input data is quantized based on the predicted minimum value and predicted maximum value. The output of the convolution is computed into an accumulator and re-quantized. The re-quantized value is output to an external memory. The predicted min value and the predicted max value are computed based on the previous max values and min values with a weighted average or a pre-determined formula. Initial min value and max value are computed based on known quantization methods and utilized for initializing the predicted min value and predicted max value in the quantization process.

A method of indexing and searching for video content is provided. For each frame of a first plurality of frames, a first global feature and a first plurality of local features may be identified. The first plurality of local features may be clustered around a first plurality of cluster centers. The first plurality of local features may be converted into a first plurality of binary signatures. An index that maps the first plurality of cluster centers and the first plurality of binary signatures to the first plurality of frames may be generated. A search request associated with a second video may be received and its direct and indirect features may be identified. The identified features of the second video may be compared against the index and a candidate video may be selected as a result of the search request.

The present invention discloses a system for automated vehicles license plates characters segmentation and recognition comprising an imaging processor connected to at least one image grabber module or camera. The image grabber module captures images of the vehicles and forwards it to said connected imaging processor and the imaging processor segments and recognizes the vehicles license plates character region including the region with deformed license plates characters in the captured vehicle images by involving binarization of maximally stable external regions corresponding to probable license plate region in the captured vehicle images.

Systems, methods, tangible non-transitory computer-readable media, and devices for autonomous vehicle operation are provided. For example, a computing system can receive object data that includes portions of sensor data. The computing system can determine, in a first stage of a multiple stage classification using hardware components, one or more first stage characteristics of the portions of sensor data based on a first machine-learned model. In a second stage of the multiple stage classification, the computing system can determine second stage characteristics of the portions of sensor data based on a second machine-learned model. The computing system can generate an object output based on the first stage characteristics and the second stage characteristics. The object output can include indications associated with detection of objects in the portions of sensor data.

Embodiments of this disclosure provide a lane detection apparatus and method and an electronic device. First, preprocessing is performed based on semantic segmentation to remove interference objects in a binary image, which may improve accuracy of a lane line detection and may be applicable to various road scenarios, and the result based on semantic segmentation may automatically extract a lane line region image containing one or more lane lines, thereby automatically performing perspective transformation to perform search and fitting on the one or more lane lines, including achieving multi-lane detection. And furthermore, by synthesizing detection results of a plurality of input images, accuracy and integrity of the lane detection may further be improved.

A method for optically inspecting a mold (10) for manufacturing ophthalmic lenses such as contact lenses for possible mold defects, including: generating a set of images of the mold (10) for different azimuthal illumination angles (1, 9) using an illumination system (20) and an imaging system (30), the latter being aligned such that its focal plane cuts through the mold (10) at a specific axial position along a center axis of the mold (10); generating a focal plane image by averaging pixelwise over the set of images after having masked out in each image those regions that include direct specular reflections from the mold (10); repeating the previous steps for one or a plurality of different axial positions of the focal plane such as to generate a plurality of different focal plane images; identifying one or more image features in the plurality of focal plane images indicative for a possible mold defect; determining for each identified image feature in which focal plane image the identified image feature appears sharpest; generating for each identified image feature a respective image section out of the respective sharpest focal plane containing the image feature; and generating a composed dark field image of the mold (10) by composing the respective image sections for each identified image feature, thus enabling to determine as to whether the possible defects of the mold (10) still allow the mold (10) to be used.

A method of locating content portion(s) of an image includes automatically determining a pixel threshold value; comparing the threshold value to the image to determine background and content sets of the pixels; determining boundaries between the background and content sets; determining the content portions using the determined boundaries; and determining bounding boxes for the content portions. A computed-radiography scanner includes an imager; a transport configured to operatively arrange a plurality of plates with respect to the imager so that a single image of the plates can be captured; and a processor configured to automatically locate the content portion(s) in the single image. A locating method includes automatically determining the pixel threshold value; determining the boundaries between portions delimited by the pixel threshold value; determining the content portions using the boundaries; and determining the bounding boxes.