Deng Linger, Linghao Zhu, Yuliang Liu, Yu Wang, Qunyi Xie, Jingjing Wu, Gang Zhang, Yingying Zhu, Xiang Bai. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Scoring the Optical Character Recognition (OCR) capabilities of Large Multimodal Models (LMMs) has witnessed growing interest. Existing benchmarks have highlighted the impressive performance of LMMs in text recognition; however, their abilities in certain challenging tasks, such as text localization, handwritten content extraction, and logical reasoning, remain underexplored. To bridge this gap, we introduce OCRBench v2, a large-scale bilingual text-centric benchmark with currently the most comprehensive set of tasks ($4\times$ more tasks than the previous multi-scene benchmark OCRBench), the widest coverage of scenarios ($31$ diverse scenarios), and thorough evaluation metrics, with $10,000$ human-verified question-answering pairs and a high proportion of difficult samples. Moreover, we construct a private test set with $1,500$ manually annotated images. The consistent evaluation trends observed across both public and private test sets validate the OCRBench v2's reliability. After carefully benchmarking state-of-the-art LMMs, we find that most LMMs score below $50$ ($100$ in total) and suffer from five-type limitations, including less frequently encountered text recognition, fine-grained perception, layout perception, complex element parsing, and logical reasoning. The benchmark and evaluation scripts are available at https://github.com/Yuliang-Liu/MultimodalOCR.

As electronic textiles have become more advanced in sensing, actuating, and manufacturing, incorporating smartness into fabrics has become of special interest to ubiquitous computing and interaction researchers and designers. However, innovating smart textile interfaces for numerous input and output modalities usually requires expert-level knowledge of specific materials, fabrication, and protocols. This paper presents EmTex, a construction kit based on embroidered textiles, patterned with dedicated sensing, actuating, and connecting components to facilitate the design and prototyping of smart textile interfaces. With machine embroidery, EmTex is compatible with a wide range of threads and underlay fabrics, proficient in various stitches to control the electric parameters, and capable of integrating versatile and reliable interaction functionalities with aesthetic patterns and precise designs. EmTex consists of 28 textile-based sensors, actuators, connectors, and displays, presented with standardized visual and tactile effects. Along with a visual programming tool, EmTex enables the prototyping of everyday textile interfaces for diverse life-living scenarios, that embody their touch input, and visual and haptic output properties. With EmTex, we conducted a workshop and invited 25 designers and makers to create freeform textile interfaces. Our findings revealed that EmTex helped the participants explore novel interaction opportunities with various smart textile prototypes. We also identified challenges EmTex shall face for practical use in promoting the design innovation of smart textiles.

ABSTRACT In this paper, the design of a low‐cost ultra‐high‐frequency (UHF) Radio Frequency IDentification (RFID) tag chip with an advanced encryption standard (AES) cryptographic engine is presented. The design of digital baseband is verified on a Field‐Programmable Gate Array (FPGA) platform. The whole chip, including a radio frequency frontend, an analog frontend, an Electrically Erasable Programmable Read‐Only Memory (EEPROM), and a baseband with AES engine, is taped out on Semiconductor Manufacturing International Corporation (SMIC) 0.13μm process. The chip area is 1 × 1 mm2, in which 0.6 × 0.3 mm2 is covered by the digital baseband. The power consumption of the entire tag chip is 20.9 μW. The design can work on both two modes of the standard ISO 18000‐6C mode and the security enhanced ISO 18000‐6C mode. To the best of our knowledge, it is the first UHF passive RFID tag chip with AES algorithm in the baseband.