Multi-class tissue-type classification of colorectal cancer (CRC) histopathologic images is a significant step in the development of downstream machine learning models for diagnosis and treatment planning. However, publicly available CRC datasets used to build tissue classifiers often suffer from insufficient morphologic diversity, class imbalance, and low-quality image tiles, limiting downstream model performance and generalizability. To address this research gap, we introduce STARC-9 (STAnford coloRectal Cancer), a large-scale dataset for multi-class tissue classification. STARC-9 comprises 630,000 histopathologic image tiles uniformly sampled across nine clinically relevant tissue classes (each represented by 70,000 tiles), systematically extracted from hematoxylin & eosin-stained whole-slide images (WSI) from 200 CRC patients at the Stanford University School of Medicine. To construct STARC-9, we propose a novel framework, DeepCluster++, consisting of two primary steps to ensure diversity within each tissue class, followed by pathologist verification. First, an encoder from an autoencoder trained specifically on histopathologic images is used to extract feature vectors from all tiles within a given input WSI. Next, K-means clustering groups morphologically similar tiles, followed by an equal-frequency binning method to sample diverse patterns within each tissue class. Finally, the selected tiles are verified by expert gastrointestinal pathologists to ensure classification accuracy. This semi-automated approach significantly reduces the manual effort required for dataset curation while producing high-quality training examples. To validate the utility of STARC-9, we benchmarked baseline convolutional neural networks, transformers, and pathology-specific foundation models on downstream multi-class CRC tissue classification and segmentation tasks when trained on STARC-9 versus publicly available datasets, demonstrating superior generalizability of models trained on STARC-9. Although we demonstrate the utility of DeepCluster++ on CRC as a pilot use-case, it is a flexible framework that can be used for constructing high-quality datasets from large WSI repositories across a wide range of cancer and non-cancer applications.

This paper tries to raise awareness of security and privacy issues associated with Integrated Sensing and Communications (ISAC). In the context of ISAC, sensing (or radar) and communication functionalities can work synergically, i.e., the two functionalities can benefit each other, which is both exciting and worrisome. Our concern is that ISAC exposes increased vulnerability and faces threats we never experienced before, while it offers tremendous opportunity. In the ISAC system not only can an attacker wirelessly sniff data transmitted by legitimate users, but can also locate them passively or actively via illuminating. For instance, sensing-assisted eavesdropping in ISAC is much easier than traditional eavesdropping. With the availability of precise sensing at sub-centimeter accuracy, an attacker gains much more information than just a MAC address. The stolen information can include location, speed and ambient condition, and may be in the format of images. Consequently, attacks can be more targeted and effective. In addition, as sensing becomes ubiquitous and collaborative sensing becomes easier, it is harder to preserve privacy and to manage the trustworthiness of a large number of participants. We introduce a secure ISAC (S-ISAC) framework and propose a number of solutions. In contrast to some information-theoretic works on ISAC security, we emphasize practical countermeasures from an ISAC-network perspective. We also propose a number of research topics that need to be addressed before the ISAC concept is fully adopted by the industry.

Electric Vehicle (EV) dynamic charging is a new charging technology that has received attention recently from both industry and academia. It is much convenient compared to plug-in power charging. However, the EVs need to pay and authenticate to charging pads placed on roads to get charged. because of its wireless and fast moving nature, dynamically offering services (including secure and reliable payment based access, power charging, etc.) to the paid users is challenging. In an efficient physical-layer-based authentication scheme, the EVs should sent digital secret token to the pads that should verify them. To design a scheme that meets both security and reliability requirements in EV dynamic charging scenario, this paper analyzes and evaluates the verification performance. Probabilistic performance metrics are defined and an analytic framework is developed to quantify the performance. The derived estimation formulas are validated by numerical results, and these results reveal the sensitive impacts of token-pool Hamming distance, decision threshold and Signal-to-Noise Ratio (SNR).