Abstract: Disclosed are a method and apparatus for measuring a gaze difference based on gaze estimation. Video data of a testee gazing at a visual target is obtained, and a face image sequence is obtained from the video data. The face image sequence is input into a first neural network for key frame extraction to obtain a first and second gaze position face image. The face images are input into a second neural network for facial feature point extraction. Cropping is performed based on facial feature point coordinates to obtain a first and second eye region image. A gaze difference estimation network is trained to obtain a trained gaze difference estimation network. The first and second eye region images are input into the trained gaze difference estimation network to obtain a predicted gaze difference between the first and second eye region images.

Abstract: Disclosed are a method and apparatus for measuring a gaze difference based on gaze estimation. Video data of a testee gazing at a visual target is obtained, and a face image sequence is obtained from the video data. The face image sequence is input into a first neural network for key frame extraction to obtain a first and second gaze position face image. The face images are input into a second neural network for facial feature point extraction. Cropping is performed based on facial feature point coordinates to obtain a first and second eye region image. A gaze difference estimation network is trained to obtain a trained gaze difference estimation network. The first and second eye region images are input into the trained gaze difference estimation network to obtain a predicted gaze difference between the first and second eye region images.

Abstract: Disclosed is a method and apparatus for measuring a deviation angle of a gaze position based on 3D reconstruction. A face image sequence of a testee is acquired. Covering conditions of the face image sequence are obtained. Key frame images are determined and input into a second neural network to obtain feature point heat maps that are converted into facial feature point coordinates. An objective function is constructed. A head pose corresponding to the key frame images is obtained. An eyeball position is initialized. An eyeball rotation angle of a reference gaze position is set as a preset angle. 3D coordinates of an eyeball in a reference gaze position image in a head coordinate system is solved. 3D coordinates of an eyeball in a to-be-measured image in the head coordinate system are fixed. An eyeball rotation angle is solved. A deviation angle of a gaze position is obtained.

Abstract: A method and an apparatus for protecting privacy of an ophthalmic patient and a storage medium are disclosed. The method includes: acquiring an examination video of the ophthalmic patient; extracting image features of each frame of images in the examination video, performing region division on each frame of the images according to position information of organs in the image features, and obtaining a region to be precisely reconstructed and a region to be roughly reconstructed according to a region division result; performing three-dimensional (3D) reconstruction on the region to be precisely reconstructed and the region to be roughly reconstructed, to obtain 3D reconstruction data corresponding to each frame of the images; and rendering all the 3D reconstruction data into a 3D reconstruction video.

Abstract: A method and an apparatus for protecting privacy of an ophthalmic patient and a storage medium are disclosed. The method includes: acquiring an examination video of the ophthalmic patient; extracting image features of each frame of images in the examination video, performing region division on each frame of the images according to position information of organs in the image features, and obtaining a region to be precisely reconstructed and a region to be roughly reconstructed according to a region division result; performing three-dimensional (3D) reconstruction on the region to be precisely reconstructed and the region to be roughly reconstructed, to obtain 3D reconstruction data corresponding to each frame of the images; and rendering all the 3D reconstruction data into a 3D reconstruction video.