Self-adaptive image segmentation method for two-photon calcium imaging video data

Abstract

The invention discloses a self-adaptive image segmentation method for two-photon calcium imaging video data. The self-adaptive image segmentation method includes: selecting the kth frame to the nth frame from the two-photon calcium imaging video to construct a training sample; constructing an initialized single-mode background model according to the training sample; continuously updating the single-mode background model in real time; and carrying out segmentation detection on the image input in real time by utilizing the single-mode background model updated in real time. According to the method, the problem that background modeling and image segmentation methods are specially designed for the characteristics of the two-photon calcium imaging video data is solved, the problem that some existing methods cannot adapt to and utilize the characteristics of the two-photon calcium imaging video data is solved, and the processing accuracy and the operation efficiency are effectively guaranteed.

Description

technical field [0001] The invention relates to an image processing method in the technical field of video data mining, in particular to an adaptive image segmentation method for two-photon calcium imaging video data. Background technique [0002] Two-photon calcium imaging videos are typically characterized by high resolution, high frame rate, and high bit depth. Take the mouse brain nerve two-photon calcium imaging video provided by the Allen Institute for Brain Science as an example. The total number of frames in a single video exceeds 100,000 frames, and the image resolution of each frame is 512×512 pixels. The data capacity of the video is More than 56GB. With such a huge amount of data, coupled with the large number of brain neurons and complex activity patterns recorded in the video, it is impossible to use manual methods for data mining. Therefore, it is necessary to study and design a set of efficient automated data mining solutions. [0003] Adaptive image segmen...

AI Technical Summary

Problems solved by technology

[0005] One type of method is derived from the traditional image segmentation method for a single static image. The problem with this method is that the video image sequence with time coherence is treated as an isolated and unrelated single image, and only a single image is used. The spatial information inside the image, but the time dimension information of the video image sequence is completely lost, so it is impossible to fully exploit and utilize the implicit dynamic information of brain neurons in the two-photon calcium imaging video

[0006] Another type of method is the image segmentation method derived from the traditional intelligent video surveillance field. The problem of these methods is: the lack of adaptation and utilization of the characteristics of two-photon calcium imaging video data

Most image segmentation methods based on 8-bit data depth either cannot handle 16-bit depth data at all, or there will be a problem of collapse in computing performance

In addition, multimodal background model frameworks are usually used in intelligent video image segmentation methods, and the background models of the above-mentioned two-photon calcium imaging videos are all single-modal. loss, and may also result in decreased detection sensitivity for active neurons

[0007] To sum up, if a mismatched adaptive image segmentation method is blindly transplanted, it will not only be unable to truly and effectively realize the data mining of two-photon calcium imaging video, but also seriously cause misjudgment of the experimental results.

Images