白丝美女被狂躁免费视频网站,500av导航大全精品,yw.193.cnc爆乳尤物未满,97se亚洲综合色区,аⅴ天堂中文在线网官网

Systems and methods for encoding image features of high-resolution digital images of biological specimens

專(zhuān)利號(hào)
US11176412B2
公開(kāi)日期
2021-11-16
申請(qǐng)人
Ventana Medical Systems, Inc.(US AZ Tucson)
發(fā)明人
Yao Nie
IPC分類(lèi)
G06K9/00; G06K9/62; G06T7/11; G06K9/46
技術(shù)領(lǐng)域
superpixel,image,clustering,centroid,clusters,in,cluster,vectors,vector,biological
地域: AZ AZ Tucson

摘要

An image analysis system for analyzing biological specimen images is disclosed. The system may include: a superpixel generator configured to obtain a biological specimen image and group pixels of the biological specimen image into a plurality of superpixels; a feature extractor configured to extract, from each superpixel in the plurality of superpixels, a feature vector comprising a plurality of image features; a clustering engine configured to assign the plurality of superpixels to a predefined number of clusters, each cluster being characterized by a centroid vector of feature vectors of superpixels assigned to the cluster; and a storage interface configured to store, for each superpixel in the plurality of superpixels, clustering information identifying the one cluster to which the superpixel is assigned. The system may also include a graph engine configured construct a graph based on the stored information, and use the graph to perform a graph-based image processing task.

說(shuō)明書(shū)

All information stored by storage interface 113 can later be retrieved by storage interface 113 and provided to image analysis system 100 that can perform various image processing tasks, examples of which are provided below. It should be evident to a person skilled in the art that by storing only limited amount of data for each superpixel, such as storing its clustering information without storing its feature vector containing its image features, significant reductions in memory consumption and in storage/retrieval times can be achieved. To illustrate with an example, let N be the number of superpixels generated for a given image; M be the number of image features extracted for each superpixel; U be the number of bytes representing each image feature; and K be the number of clusters. Thus, the original feature vectors for all superpixels occupy N×M×U bytes. In contrast, it would take only N×U bytes to store the clustering information for each superpixel, K×M×U bytes to store the centroid vectors of all clusters, and K*(K?1)/2×U bytes to store the distances between each two clusters. Thus, using the techniques described herein, a data compression ratio of at least N×M/(N+K×M+K×(K?1)/2) can be achieved if centroid vectors are stored, and a compression ratio of at least N×M/(N+K×(K?1)/2) can be achieved if centroid vectors are not stored.

權(quán)利要求

1
微信群二維碼
意見(jiàn)反饋