This paper proposes a novel noise-shaping (NS) successive approximation register (SAR) based light-to-digital converter (LDC) for continuous multi-wavelength (MW) photoplethysmogram (PPG) monitoring. The proposed NS-SAR LDC employs a single-channel current-integration (CI) SAR quantizer with a three-channel passive NS loop filter performs time-multiplexed, low-power NS readout for MWPPG. The cascade integrator feedforward NS architecture features a passive nested feedback path, which contributes one additional NS order. The LDC enables a wide dynamic range (DR) through the high signal-to-noise-and-distortion ratio (SNDR), inherent AC gain control of the proposed CI NS-SAR, and a predictive threshold filter that compensates for large DC and drifting baseline currents induced by motion artifacts and interference light. Fabricated in a 180-nm standard CMOS process, the NS-SAR LDC consumes$26.5~\mu $W while achieving a SNDR of 95 dB and a total DR of 161.5 dB. The work is validated through finger MWPPG measurements using commercial PPG sensors.

Multimodal entity linking plays a crucial role in a wide range of applications. Recent advances in large language model-based methods have become the dominant paradigm for this task, effectively leveraging both textual and visual modalities to enhance performance. Despite their success, these methods still face two challenges, including unnecessary incorporation of image data in certain scenarios and the reliance only on a one-time extraction of visual features, which can undermine their effectiveness and accuracy. To address these challenges, we propose a novel LLM-based framework for the multimodal entity linking task, called Intra- and Inter-modal Collaborative Reflections. This framework prioritizes leveraging text information to address the task. When text alone is insufficient to link the correct entity through intra- and inter-modality evaluations, it employs a multi-round iterative strategy that integrates key visual clues from various aspects of the image to support reasoning and enhance matching accuracy. Extensive experiments on three widely used public datasets demonstrate that our framework consistently outperforms current state-of-the-art methods in the task, achieving improvements of 3.2%, 5.1%, and 1.6%, respectively. Our code is available at https://github.com/ziyan-xiaoyu/I2CR/.

This work presents a low-power, large cross-scale dynamic range (DR) successive approximation register (SAR) based light-to-digital converter (LDC) designed for wearable multi-wavelength photoplethysmogram (MWPPG) readout. The proposed SAR-LDC employs a current-integration SAR (CI-SAR) quantizer that directly integrates the photocurrent onto a binary capacitive digital-to-analog converter (DAC) and utilizes low-power SAR logic for direct digitization. This approach releases the various trade-offs present in the recent LDC designs and further maximizes the LDC’s power efficiency. At the circuit level, an AC/DC correlated DR enhancement loop, controlled by a digital reconfigurable unit (DRU), enhances both the DC and AC DR of the proposed system. The proposed architecture is fabricated in standard 180-nm CMOS technology. By utilizing a 10-bit SAR-LDC, it achieves a maximum effective number of bits (ENOB) of 9.93 and 12.56 in the Nyquist bandwidth and PPG bandwidth, respectively. The SAR-LDC reads MWPPG on a single channel through time multiplexing, consuming 9.84 μW only while achieving a cross-scale AC DR of 103.4 dB and a total DR of 148.9 dB without using any hybrid circuit technique. The proposed SAR-LDC is further validated using with a commercial PPG sensor for finger MWPPG measurement.