ObjectFinder: Open-Vocabulary Assistive System for Interactive Object Search by Blind People

Assistive technology can be leveraged by blind people when searching for objects in their daily lives. We created ObjectFinder, an open-vocabulary interactive object-search prototype, which combines object detection with scene description and navigation. It enables blind persons to detect and navigate to objects of their choice. Our approach used co-design for the development of the prototype. We further conducted need-finding interviews to better understand challenges in object search, followed by a study with the ObjectFinder prototype in a laboratory setting simulating a living room and an office, with eight blind users. Additionally, we compared the prototype with BeMyEyes and Lookout for object search. We found that most participants felt more independent with ObjectFinder and preferred it over the baselines when deployed on more efficient hardware, as it enhances mental mapping and allows for active target definition. Moreover, we identified factors for future directions for the development of object-search systems.

Deformable Mamba for Wide Field of View Segmentation

Wide-FoV cameras, like fisheye and panoramic setups, are essential for broader perception but introduce significant distortions in 180{\deg} and 360{\deg} images, complicating dense prediction tasks. For instance, existing MAMBA models lacking distortion-aware capacity cannot perform well in panoramic semantic segmentation. To address this problem, this work presents Deformable Mamba, a unified framework specifically designed to address imaging distortions within the context of panoramic and fisheye semantic segmentation. At the core is a decoder constructed with a series of Deformable Mamba Fusion (DMF) blocks, making the whole framework more deformable, efficient, and accurate, when handling extreme distortions. Extensive evaluations across five datasets demonstrate that our method consistently improves segmentation accuracy compared to the previous state-of-the-art methods tailored for specific FoVs. Notably, Deformable Mamba achieves a +2.5% performance improvement on the 360{\deg} Stanford2D3D dataset, and shows better results across FoVs from 60{\deg} to 360{\deg}.

Solving Zero-Shot 3D Visual Grounding as Constraint Satisfaction Problems

3D visual grounding (3DVG) aims to locate objects in a 3D scene with natural language descriptions. Supervised methods have achieved decent accuracy, but have a closed vocabulary and limited language understanding ability. Zero-shot methods mostly utilize large language models (LLMs) to handle natural language descriptions, yet suffer from slow inference speed. To address these problems, in this work, we propose a zero-shot method that reformulates the 3DVG task as a Constraint Satisfaction Problem (CSP), where the variables and constraints represent objects and their spatial relations, respectively. This allows a global reasoning of all relevant objects, producing grounding results of both the target and anchor objects. Moreover, we demonstrate the flexibility of our framework by handling negation- and counting-based queries with only minor extra coding efforts. Our system, Constraint Satisfaction Visual Grounding (CSVG), has been extensively evaluated on the public datasets ScanRefer and Nr3D datasets using only open-source LLMs. Results show the effectiveness of CSVG and superior grounding accuracy over current state-of-the-art zero-shot 3DVG methods with improvements of $+7.0\%$ (Acc@0.5 score) and $+11.2\%$ on the ScanRefer and Nr3D datasets, respectively. The code of our system is publicly available at https://github.com/sunsleaf/CSVG.

TAPT: Test-Time Adversarial Prompt Tuning for Robust Inference in Vision-Language Models

Large pre-trained Vision-Language Models (VLMs) such as CLIP have demonstrated excellent zero-shot generalizability across various downstream tasks. However, recent studies have shown that the inference performance of CLIP can be greatly degraded by small adversarial perturbations, especially its visual modality, posing significant safety threats. To mitigate this vulnerability, in this paper, we propose a novel defense method called Test-Time Adversarial Prompt Tuning (TAPT) to enhance the inference robustness of CLIP against visual adversarial attacks. TAPT is a test-time defense method that learns defensive bimodal (textual and visual) prompts to robustify the inference process of CLIP. Specifically, it is an unsupervised method that optimizes the defensive prompts for each test sample by minimizing a multi-view entropy and aligning adversarial-clean distributions. We evaluate the effectiveness of TAPT on 11 benchmark datasets, including ImageNet and 10 other zero-shot datasets, demonstrating that it enhances the zero-shot adversarial robustness of the original CLIP by at least 48.9% against AutoAttack (AA), while largely maintaining performance on clean examples. Moreover, TAPT outperforms existing adversarial prompt tuning methods across various backbones, achieving an average robustness improvement of at least 36.6%.

Expose Before You Defend: Unifying and Enhancing Backdoor Defenses via Exposed Models

Backdoor attacks covertly implant triggers into deep neural networks (DNNs) by poisoning a small portion of the training data with pre-designed backdoor triggers. This vulnerability is exacerbated in the era of large models, where extensive (pre-)training on web-crawled datasets is susceptible to compromise. In this paper, we introduce a novel two-step defense framework named Expose Before You Defend (EBYD). EBYD unifies existing backdoor defense methods into a comprehensive defense system with enhanced performance. Specifically, EBYD first exposes the backdoor functionality in the backdoored model through a model preprocessing step called backdoor exposure, and then applies detection and removal methods to the exposed model to identify and eliminate the backdoor features. In the first step of backdoor exposure, we propose a novel technique called Clean Unlearning (CUL), which proactively unlearns clean features from the backdoored model to reveal the hidden backdoor features. We also explore various model editing/modification techniques for backdoor exposure, including fine-tuning, model sparsification, and weight perturbation. Using EBYD, we conduct extensive experiments on 10 image attacks and 6 text attacks across 2 vision datasets (CIFAR-10 and an ImageNet subset) and 4 language datasets (SST-2, IMDB, Twitter, and AG's News). The results demonstrate the importance of backdoor exposure for backdoor defense, showing that the exposed models can significantly benefit a range of downstream defense tasks, including backdoor label detection, backdoor trigger recovery, backdoor model detection, and backdoor removal. We hope our work could inspire more research in developing advanced defense frameworks with exposed models. Our code is available at: https://github.com/bboylyg/Expose-Before-You-Defend.

An Image-Guided Robotic System for Transcranial Magnetic Stimulation: System Development and Experimental Evaluation

Transcranial magnetic stimulation (TMS) is a noninvasive medical procedure that can modulate brain activity, and it is widely used in neuroscience and neurology research. Compared to manual operators, robots may improve the outcome of TMS due to their superior accuracy and repeatability. However, there has not been a widely accepted standard protocol for performing robotic TMS using fine-segmented brain images, resulting in arbitrary planned angles with respect to the true boundaries of the modulated cortex. Given that the recent study in TMS simulation suggests a noticeable difference in outcomes when using different anatomical details, cortical shape should play a more significant role in deciding the optimal TMS coil pose. In this work, we introduce an image-guided robotic system for TMS that focuses on (1) establishing standardized planning methods and heuristics to define a reference (true zero) for the coil poses and (2) solving the issue that the manual coil placement requires expert hand-eye coordination which often leading to low repeatability of the experiments. To validate the design of our robotic system, a phantom study and a preliminary human subject study were performed. Our results show that the robotic method can half the positional error and improve the rotational accuracy by up to two orders of magnitude. The accuracy is proven to be repeatable because the standard deviation of multiple trials is lowered by an order of magnitude. The improved actuation accuracy successfully translates to the TMS application, with a higher and more stable induced voltage in magnetic field sensors.

Debiasing Vison-Language Models with Text-Only Training

Pre-trained vision-language models (VLMs), such as CLIP, have exhibited remarkable performance across various downstream tasks by aligning text and images in a unified embedding space. However, due to the imbalanced distribution of pre-trained datasets, CLIP suffers from the bias problem in real-world applications. Existing debiasing methods struggle to obtain sufficient image samples for minority groups and incur high costs for group labeling. To address the limitations, we propose a Text-Only Debiasing framework called TOD, leveraging a text-as-image training paradigm to mitigate visual biases. Specifically, this approach repurposes the text encoder to function as an image encoder, thereby eliminating the need for image data. Simultaneously, it utilizes a large language model (LLM) to generate a balanced text dataset, which is then used for prompt tuning. However, we observed that the model overfits to the text modality because label names, serving as supervision signals, appear explicitly in the texts. To address this issue, we further introduce a Multi-Target Prediction (MTP) task that motivates the model to focus on complex contexts and distinguish between target and biased information. Extensive experiments on the Waterbirds and CelebA datasets show that our method significantly improves group robustness, achieving state-of-the-art results among image-free methods and even competitive performance compared to image-supervised methods. Furthermore, the proposed method can be adapted to challenging scenarios with multiple or unknown bias attributes, demonstrating its strong generalization and robustness.

AnyAttack: Towards Large-scale Self-supervised Generation of Targeted Adversarial Examples for Vision-Language Models

Due to their multimodal capabilities, Vision-Language Models (VLMs) have found numerous impactful applications in real-world scenarios. However, recent studies have revealed that VLMs are vulnerable to image-based adversarial attacks, particularly targeted adversarial images that manipulate the model to generate harmful content specified by the adversary. Current attack methods rely on predefined target labels to create targeted adversarial attacks, which limits their scalability and applicability for large-scale robustness evaluations. In this paper, we propose AnyAttack, a self-supervised framework that generates targeted adversarial images for VLMs without label supervision, allowing any image to serve as a target for the attack. To address the limitation of existing methods that require label supervision, we introduce a contrastive loss that trains a generator on a large-scale unlabeled image dataset, LAION-400M dataset, for generating targeted adversarial noise. This large-scale pre-training endows our method with powerful transferability across a wide range of VLMs. Extensive experiments on five mainstream open-source VLMs (CLIP, BLIP, BLIP2, InstructBLIP, and MiniGPT-4) across three multimodal tasks (image-text retrieval, multimodal classification, and image captioning) demonstrate the effectiveness of our attack. Additionally, we successfully transfer AnyAttack to multiple commercial VLMs, including Google's Gemini, Claude's Sonnet, and Microsoft's Copilot. These results reveal an unprecedented risk to VLMs, highlighting the need for effective countermeasures.

@Bench: Benchmarking Vision-Language Models for Human-centered Assistive Technology

As Vision-Language Models (VLMs) advance, human-centered Assistive Technologies (ATs) for helping People with Visual Impairments (PVIs) are evolving into generalists, capable of performing multiple tasks simultaneously. However, benchmarking VLMs for ATs remains under-explored. To bridge this gap, we first create a novel AT benchmark (@Bench). Guided by a pre-design user study with PVIs, our benchmark includes the five most crucial vision-language tasks: Panoptic Segmentation, Depth Estimation, Optical Character Recognition (OCR), Image Captioning, and Visual Question Answering (VQA). Besides, we propose a novel AT model (@Model) that addresses all tasks simultaneously and can be expanded to more assistive functions for helping PVIs. Our framework exhibits outstanding performance across tasks by integrating multi-modal information, and it offers PVIs a more comprehensive assistance. Extensive experiments prove the effectiveness and generalizability of our framework.

OneBEV: Using One Panoramic Image for Bird's-Eye-View Semantic Mapping

In the field of autonomous driving, Bird's-Eye-View (BEV) perception has attracted increasing attention in the community since it provides more comprehensive information compared with pinhole front-view images and panoramas. Traditional BEV methods, which rely on multiple narrow-field cameras and complex pose estimations, often face calibration and synchronization issues. To break the wall of the aforementioned challenges, in this work, we introduce OneBEV, a novel BEV semantic mapping approach using merely a single panoramic image as input, simplifying the mapping process and reducing computational complexities. A distortion-aware module termed Mamba View Transformation (MVT) is specifically designed to handle the spatial distortions in panoramas, transforming front-view features into BEV features without leveraging traditional attention mechanisms. Apart from the efficient framework, we contribute two datasets, i.e., nuScenes-360 and DeepAccident-360, tailored for the OneBEV task. Experimental results showcase that OneBEV achieves state-of-the-art performance with 51.1% and 36.1% mIoU on nuScenes-360 and DeepAccident-360, respectively. This work advances BEV semantic mapping in autonomous driving, paving the way for more advanced and reliable autonomous systems.