Faster Mean-shift: GPU-accelerated clustering for cosine embedding-based cell segmentation and tracking

Abstract

Recently, single-stage embedding based deep learning algorithms gain increasing attention in cell segmentation and tracking. Compared with the traditional "segment-then-associate" two-stage approach, a single-stage algorithm not only simultaneously achieves consistent instance cell segmentation and tracking but also gains superior performance when distinguishing ambiguous pixels on boundaries and overlaps. However, the deployment of an embedding based algorithm is restricted by slow inference speed (e.g., ≈1-2 mins per frame). In this study, we propose a novel Faster Mean-shift algorithm, which tackles the computational bottleneck of embedding based cell segmentation and tracking. Different from previous GPU-accelerated fast mean-shift algorithms, a new online seed optimization policy (OSOP) is introduced to adaptively determine the minimal number of seeds, accelerate computation, and save GPU memory. With both embedding simulation and empirical validation via the four cohorts from the ISBI cell tracking challenge, the proposed Faster Mean-shift algorithm achieved 7-10 times speedup compared to the state-of-the-art embedding based cell instance segmentation and tracking algorithm. Our Faster Mean-shift algorithm also achieved the highest computational speed compared to other GPU benchmarks with optimized memory consumption. The Faster Meanshift is a plug-and-play model, which can be employed on other pixel embedding based clustering inference for medical image analysis.

VoxelEmbed: 3D Instance Segmentation and Tracking with Voxel Embedding based Deep Learning

“…The standard Mean-shift clustering algorithm was used in Payer et al [25]. In this study, we utilize a GPU accelerated Faster Mean-shift algorithm [29] to accelerate the embedding clustering. From the unsupervised clustering, a unique label will be assigned to each cell as different instances, which achieve the instance segmentation and tracking simultaneously.…”

ASIST: Annotation-free Synthetic Instance Segmentation and Tracking by Adversarial Simulations

“…The ideal pixel-embedding has two properties: (1) embedding of pixels belonging to the same objects should be similar across the entire video, and (2) the embedding of pixels belonging to different objects should be different. For a testing video, we employed the Faster Mean-shift algorithm [6] to cluster pixels to objects as the instance segmentation and tracking results. The embeddingbased deep learning methods approach the instance segmentation and tracking as a "single-stage" approach, which is a simple and generalizable solution across different applications [7], [6].…”

“…For a testing video, we employed the Faster Mean-shift algorithm [6] to cluster pixels to objects as the instance segmentation and tracking results. The embeddingbased deep learning methods approach the instance segmentation and tracking as a "single-stage" approach, which is a simple and generalizable solution across different applications [7], [6].…”

“…Most of the previous solutions [3], [4], [5] follow a similar "two-stage" strategy: (1) segmentation on each frame, and (2) frame-by-frame association across the video. In recent years, a new family of "single-stage" algorithms was enabled by cutting-edge pixel-embedding based deep learning [6], [7]. Such methods enforce the spatio-temporally consistent pixelwise feature embedding for the same cellular or subcellular objects across video frames, which address both the instance segmentation and tracking as a holistic model.…”

How to Test the Randomness from the Wireless Channel for Security?

“…[51] proposes physical-layer authentication schemes through adding low-power signal. [52] solves the authentication in IoT by exploiting the fading of the wireless links between devices to be authenticated and a set of trusted anchor nodes. [17] proposes a retroactive key setup to protect source authentication and path validation into the realm of practicality.…”