Navigational aids for blind and low vision individuals struggle conveying dynamic real-world environments, leading to cognitive overload from continuous, undifferentiated feedback. We present AMAVA, a novel real-time video-to-audio framework that converts mobile device video into contextually relevant sound effects or text-to-speech descriptions. We propose a motion-aware pipeline using a lightweight AI classification model to distinguish between low and high-movement scenes followed by a real-time text-to-audio synthesis pipeline to enhance environmental perception more efficiently. In static environments, AMAVA generates spoken audio scene descriptions for situational awareness. In high-movement situations, it prioritizes safety by delivering sound cues, such as spoken hazard alerts and environmental sound effects. These audio outputs are produced by a decoder-only transformer-based vision-language model with mixture-of-experts and cross-modal attention for visual understanding, in conjunction with neural text-to-speech and natural sound synthesis networks. The proposed framework uses prompt-based caching and category-specific throttling to avoid auditory clutter and minimize latency. We present a comprehensive evaluation of the system, including a real-time navigation study comparing a white cane alone versus with AMAVA, that shows a significant increase in user confidence and perceived safety.

Rapid progress in unconstrained face recognition has resulted in a saturation in recognition accuracy for current benchmark datasets. While important for early progress, a chief limitation in most benchmark datasets is the use of a commodity face detector to select face imagery. The implication of this strategy is restricted variations in face pose and other confounding factors. This paper introduces the IARPA Janus Benchmark A (IJB-A), a publicly available media in the wild dataset containing 500 subjects with manually localized face images. Key features of the IJB-A dataset are: (i) full pose variation, (ii) joint use for face recognition and face detection benchmarking, (iii) a mix of images and videos, (iv) wider geographic variation of subjects, (v) protocols supporting both open-set identification (1:N search) and verification (1:1 comparison), (vi) an optional protocol that allows modeling of gallery subjects, and (vii) ground truth eye and nose locations. The dataset has been developed using 1,501,267 million crowd sourced annotations. Baseline accuracies for both face detection and face recognition from commercial and open source algorithms demonstrate the challenge offered by this new unconstrained benchmark.

We present the Discriminative Ferns Ensemble (DFE) classifier for efficient visual object recognition. The classifier architecture is designed to optimize both classification speed and accuracy when a large training set is available. Speed is obtained using simple binary features and direct indexing into a set of tables, and accuracy by using a large capacity model and careful discriminative optimization. The proposed framework is applied to the problem of hand pose recognition in depth and infra-red images, using a very large training set. Both the accuracy and the classification time obtained are considerably superior to relevant competing methods, allowing one to reach accuracy targets with run times orders of magnitude faster than the competition. We show empirically that using DFE, we can significantly reduce classification time by increasing training sample size for a fixed target accuracy. Finally a DFE result is shown for the MNIST dataset, showing the method's merit extends beyond depth images.

The number of images and videos available for search on the Internet is in the order of a trillion images, making current brute force search techniques prohibitively inefficient on such large scales. As society continues to increase our desire and ability to search these vast collections of data, improving upon traditional face recognition search techniques becomes an important problem to address. Because face recognition (and other biometric) algorithms are only commercially available as black box systems, any indexing scheme developed to perform efficient search must operate without access to the underlying feature vectors used to measure facial similarity. To address this restriction, we propose a structured search that separates the facial feature space into clusters derived from sets of prototype subjects we refer to as “synecdoches”. After an offline training step, our proposed method assigns each gallery image to a cluster in the face space based on its similarity to a set of synecdoche clusters. In turn, query images are compared to the target gallery images based on the closest synecdoche cluster in sequence. Our results show a minimal drop in accuracy when only considering half of the clusters, thus reducing the search space in half. Additional experiments demonstrate the viability of our proposed approach to improve search efficiency amidst the common restriction of a black box matcher.