Yash Bhalgat

Yash Bhalgat, Iro Laina, João F. Henriques et al.

Instance segmentation in 3D is a challenging task due to the lack of large-scale annotated datasets. In this paper, we show that this task can be addressed effectively by leveraging instead 2D pre-trained models for instance segmentation. We propose a novel approach to lift 2D segments to 3D and fuse them by means of a neural field representation, which encourages multi-view consistency across frames. The core of our approach is a slow-fast clustering objective function, which is scalable and well-suited for scenes with a large number of objects. Unlike previous approaches, our method does not require an upper bound on the number of objects or object tracking across frames. To demonstrate the scalability of the slow-fast clustering, we create a new semi-realistic dataset called the Messy Rooms dataset, which features scenes with up to 500 objects per scene. Our approach outperforms the state-of-the-art on challenging scenes from the ScanNet, Hypersim, and Replica datasets, as well as on our newly created Messy Rooms dataset, demonstrating the effectiveness and scalability of our slow-fast clustering method.

Hugo Berg, Siobhan Mackenzie Hall, Yash Bhalgat et al. · oxford

Vision-language models can encode societal biases and stereotypes, but there are challenges to measuring and mitigating these multimodal harms due to lacking measurement robustness and feature degradation. To address these challenges, we investigate bias measures and apply ranking metrics for image-text representations. We then investigate debiasing methods and show that prepending learned embeddings to text queries that are jointly trained with adversarial debiasing and a contrastive loss reduces various bias measures with minimal degradation to the image-text representation.

Shuai Chen, Yash Bhalgat, Xinghui Li et al. · bytedance, oxford

Absolute Pose Regression (APR) methods use deep neural networks to directly regress camera poses from RGB images. However, the predominant APR architectures only rely on 2D operations during inference, resulting in limited accuracy of pose estimation due to the lack of 3D geometry constraints or priors. In this work, we propose a test-time refinement pipeline that leverages implicit geometric constraints using a robust feature field to enhance the ability of APR methods to use 3D information during inference. We also introduce a novel Neural Feature Synthesizer (NeFeS) model, which encodes 3D geometric features during training and directly renders dense novel view features at test time to refine APR methods. To enhance the robustness of our model, we introduce a feature fusion module and a progressive training strategy. Our proposed method achieves state-of-the-art single-image APR accuracy on indoor and outdoor datasets.

Yash Bhalgat, Joao F. Henriques, Andrew Zisserman

Transformers are powerful visual learners, in large part due to their conspicuous lack of manually-specified priors. This flexibility can be problematic in tasks that involve multiple-view geometry, due to the near-infinite possible variations in 3D shapes and viewpoints (requiring flexibility), and the precise nature of projective geometry (obeying rigid laws). To resolve this conundrum, we propose a "light touch" approach, guiding visual Transformers to learn multiple-view geometry but allowing them to break free when needed. We achieve this by using epipolar lines to guide the Transformer's cross-attention maps, penalizing attention values outside the epipolar lines and encouraging higher attention along these lines since they contain geometrically plausible matches. Unlike previous methods, our proposal does not require any camera pose information at test-time. We focus on pose-invariant object instance retrieval, where standard Transformer networks struggle, due to the large differences in viewpoint between query and retrieved images. Experimentally, our method outperforms state-of-the-art approaches at object retrieval, without needing pose information at test-time.

Yash Bhalgat, Vadim Tschernezki, Iro Laina et al.

Egocentric videos present unique challenges for 3D scene understanding due to rapid camera motion, frequent object occlusions, and limited object visibility. This paper introduces a novel approach to instance segmentation and tracking in first-person video that leverages 3D awareness to overcome these obstacles. Our method integrates scene geometry, 3D object centroid tracking, and instance segmentation to create a robust framework for analyzing dynamic egocentric scenes. By incorporating spatial and temporal cues, we achieve superior performance compared to state-of-the-art 2D approaches. Extensive evaluations on the challenging EPIC Fields dataset demonstrate significant improvements across a range of tracking and segmentation consistency metrics. Specifically, our method outperforms the next best performing approach by $7$ points in Association Accuracy (AssA) and $4.5$ points in IDF1 score, while reducing the number of ID switches by $73\%$ to $80\%$ across various object categories. Leveraging our tracked instance segmentations, we showcase downstream applications in 3D object reconstruction and amodal video object segmentation in these egocentric settings.

Xianzheng Ma, Tao Sun, Shuai Chen et al.

Recent 3D Large-Language Models (3D-LLMs) claim to understand 3D worlds, especially spatial relationships among objects. Yet, we find that simply fine-tuning a language model on text-only question-answer pairs can perform comparably or even surpass these methods on the SQA3D benchmark without using any 3D input. This indicates that the SQA3D benchmark may not be able to detect if the model exploits textual shortcuts rather than engages in 3D-aware reasoning. To address this issue, we introduce Real-3DQA, a more rigorous evaluation benchmark that filters out easy-to-guess questions and introduces a structured taxonomy to assess various aspects of 3D reasoning. Experiments on Real-3DQA confirm that existing 3D-LLMs struggle with spatial relationships once simple cues are removed. We further propose a 3D-reweighted training objective that guides model to rely more on 3D visual clues, substantially enhancing 3D-LLMs performance in spatial reasoning tasks. Our findings underscore the need for robust benchmarks and tailored training strategies to advance genuine 3D vision-language understanding. Project page: https://real-3dqa.github.io/.

Changkun Liu, Shuai Chen, Yash Bhalgat et al.

We leverage 3D Gaussian Splatting (3DGS) as a scene representation and propose a novel test-time camera pose refinement (CPR) framework, GS-CPR. This framework enhances the localization accuracy of state-of-the-art absolute pose regression and scene coordinate regression methods. The 3DGS model renders high-quality synthetic images and depth maps to facilitate the establishment of 2D-3D correspondences. GS-CPR obviates the need for training feature extractors or descriptors by operating directly on RGB images, utilizing the 3D foundation model, MASt3R, for precise 2D matching. To improve the robustness of our model in challenging outdoor environments, we incorporate an exposure-adaptive module within the 3DGS framework. Consequently, GS-CPR enables efficient one-shot pose refinement given a single RGB query and a coarse initial pose estimation. Our proposed approach surpasses leading NeRF-based optimization methods in both accuracy and runtime across indoor and outdoor visual localization benchmarks, achieving new state-of-the-art accuracy on two indoor datasets. The project page is available at https://xrim-lab.github.io/GS-CPR/.

Xianzheng Ma, Brandon Smart, Yash Bhalgat et al. · bytedance, oxford

As large language models (LLMs) evolve, their integration with 3D spatial data (3D-LLMs) has seen rapid progress, offering unprecedented capabilities for understanding and interacting with physical spaces. This survey provides a comprehensive overview of the methodologies enabling LLMs to process, understand, and generate 3D data. Highlighting the unique advantages of LLMs, such as in-context learning, step-by-step reasoning, open-vocabulary capabilities, and extensive world knowledge, we underscore their potential to significantly advance spatial comprehension and interaction within embodied Artificial Intelligence (AI) systems. Our investigation spans various 3D data representations, from point clouds to Neural Radiance Fields (NeRFs). It examines their integration with LLMs for tasks such as 3D scene understanding, captioning, question-answering, and dialogue, as well as LLM-based agents for spatial reasoning, planning, and navigation. The paper also includes a brief review of other methods that integrate 3D and language. The meta-analysis presented in this paper reveals significant progress yet underscores the necessity for novel approaches to harness the full potential of 3D-LLMs. Hence, with this paper, we aim to chart a course for future research that explores and expands the capabilities of 3D-LLMs in understanding and interacting with the complex 3D world. To support this survey, we have established a project page where papers related to our topic are organized and listed: https://github.com/ActiveVisionLab/Awesome-LLM-3D.

Ankit Dhiman, Manan Shah, Rishubh Parihar et al.

We tackle the problem of generating highly realistic and plausible mirror reflections using diffusion-based generative models. We formulate this problem as an image inpainting task, allowing for more user control over the placement of mirrors during the generation process. To enable this, we create SynMirror, a large-scale dataset of diverse synthetic scenes with objects placed in front of mirrors. SynMirror contains around 198k samples rendered from 66k unique 3D objects, along with their associated depth maps, normal maps and instance-wise segmentation masks, to capture relevant geometric properties of the scene. Using this dataset, we propose a novel depth-conditioned inpainting method called MirrorFusion, which generates high-quality, realistic, shape and appearance-aware reflections of real-world objects. MirrorFusion outperforms state-of-the-art methods on SynMirror, as demonstrated by extensive quantitative and qualitative analysis. To the best of our knowledge, we are the first to successfully tackle the challenging problem of generating controlled and faithful mirror reflections of an object in a scene using diffusion-based models. SynMirror and MirrorFusion open up new avenues for image editing and augmented reality applications for practitioners and researchers alike. The project page is available at: https://val.cds.iisc.ac.in/reflecting-reality.github.io/.

Manan Shah, Yash Bhalgat

This report is a reproducibility study of the paper "CDUL: CLIP-Driven Unsupervised Learning for Multi-Label Image Classification" (Abdelfattah et al, ICCV 2023). Our report makes the following contributions: (1) We provide a reproducible, well commented and open-sourced code implementation for the entire method specified in the original paper. (2) We try to verify the effectiveness of the novel aggregation strategy which uses the CLIP model to initialize the pseudo labels for the subsequent unsupervised multi-label image classification task. (3) We try to verify the effectiveness of the gradient-alignment training method specified in the original paper, which is used to update the network parameters and pseudo labels. The code can be found at https://github.com/cs-mshah/CDUL

Yash Bhalgat, Iro Laina, João F. Henriques et al.

Understanding complex scenes at multiple levels of abstraction remains a formidable challenge in computer vision. To address this, we introduce Nested Neural Feature Fields (N2F2), a novel approach that employs hierarchical supervision to learn a single feature field, wherein different dimensions within the same high-dimensional feature encode scene properties at varying granularities. Our method allows for a flexible definition of hierarchies, tailored to either the physical dimensions or semantics or both, thereby enabling a comprehensive and nuanced understanding of scenes. We leverage a 2D class-agnostic segmentation model to provide semantically meaningful pixel groupings at arbitrary scales in the image space, and query the CLIP vision-encoder to obtain language-aligned embeddings for each of these segments. Our proposed hierarchical supervision method then assigns different nested dimensions of the feature field to distill the CLIP embeddings using deferred volumetric rendering at varying physical scales, creating a coarse-to-fine representation. Extensive experiments show that our approach outperforms the state-of-the-art feature field distillation methods on tasks such as open-vocabulary 3D segmentation and localization, demonstrating the effectiveness of the learned nested feature field.

Yifu Tao, Yash Bhalgat, Lanke Frank Tarimo Fu et al.

We present a neural-field-based large-scale reconstruction system that fuses lidar and vision data to generate high-quality reconstructions that are geometrically accurate and capture photo-realistic textures. This system adapts the state-of-the-art neural radiance field (NeRF) representation to also incorporate lidar data which adds strong geometric constraints on the depth and surface normals. We exploit the trajectory from a real-time lidar SLAM system to bootstrap a Structure-from-Motion (SfM) procedure to both significantly reduce the computation time and to provide metric scale which is crucial for lidar depth loss. We use submapping to scale the system to large-scale environments captured over long trajectories. We demonstrate the reconstruction system with data from a multi-camera, lidar sensor suite onboard a legged robot, hand-held while scanning building scenes for 600 metres, and onboard an aerial robot surveying a multi-storey mock disaster site-building. Website: https://ori-drs.github.io/projects/silvr/

Octave Mariotti, Zhipeng Du, Yash Bhalgat et al.

Semantic correspondence (SC) aims to establish semantically meaningful matches across different instances of an object category. We illustrate how recent supervised SC methods remain limited in their ability to generalize beyond sparsely annotated training keypoints, effectively acting as keypoint detectors. To address this, we propose a novel approach for learning dense correspondences by lifting 2D keypoints into a canonical 3D space using monocular depth estimation. Our method constructs a continuous canonical manifold that captures object geometry without requiring explicit 3D supervision or camera annotations. Additionally, we introduce SPair-U, an extension of SPair-71k with novel keypoint annotations, to better assess generalization. Experiments not only demonstrate that our model significantly outperforms supervised baselines on unseen keypoints, highlighting its effectiveness in learning robust correspondences, but that unsupervised baselines outperform supervised counterparts when generalized across different datasets.

Zirui Wang, Yash Bhalgat, Ruining Li et al.

We tackle active view selection in novel view synthesis and 3D reconstruction. Existing methods like FisheRF and ActiveNeRF select the next best view by minimizing uncertainty or maximizing information gain in 3D, but they require specialized designs for different 3D representations and involve complex modelling in 3D space. Instead, we reframe this as a 2D image quality assessment (IQA) task, selecting views where current renderings have the lowest quality. Since ground-truth images for candidate views are unavailable, full-reference metrics like PSNR and SSIM are inapplicable, while no-reference metrics, such as MUSIQ and MANIQA, lack the essential multi-view context. Inspired by a recent cross-referencing quality framework CrossScore, we train a model to predict SSIM within a multi-view setup and use it to guide view selection. Our cross-reference IQA framework achieves substantial quantitative and qualitative improvements across standard benchmarks, while being agnostic to 3D representations, and runs 14-33 times faster than previous methods.

John Yang, Yash Bhalgat, Simyung Chang et al.

While hand pose estimation is a critical component of most interactive extended reality and gesture recognition systems, contemporary approaches are not optimized for computational and memory efficiency. In this paper, we propose a tiny deep neural network of which partial layers are recursively exploited for refining its previous estimations. During its iterative refinements, we employ learned gating criteria to decide whether to exit from the weight-sharing loop, allowing per-sample adaptation in our model. Our network is trained to be aware of the uncertainty in its current predictions to efficiently gate at each iteration, estimating variances after each loop for its keypoint estimates. Additionally, we investigate the effectiveness of end-to-end and progressive training protocols for our recursive structure on maximizing the model capacity. With the proposed setting, our method consistently outperforms state-of-the-art 2D/3D hand pose estimation approaches in terms of both accuracy and efficiency for widely used benchmarks.

Debasmit Das, Yash Bhalgat, Fatih Porikli

In this work, we propose a data-driven scheme to initialize the parameters of a deep neural network. This is in contrast to traditional approaches which randomly initialize parameters by sampling from transformed standard distributions. Such methods do not use the training data to produce a more informed initialization. Our method uses a sequential layer-wise approach where each layer is initialized using its input activations. The initialization is cast as an optimization problem where we minimize a combination of encoding and decoding losses of the input activations, which is further constrained by a user-defined latent code. The optimization problem is then restructured into the well-known Sylvester equation, which has fast and efficient gradient-free solutions. Our data-driven method achieves a boost in performance compared to random initialization methods, both before start of training and after training is over. We show that our proposed method is especially effective in few-shot and fine-tuning settings. We conclude this paper with analyses on time complexity and the effect of different latent codes on the recognition performance.

Yash Bhalgat, Yizhe Zhang, Jamie Lin et al.

In this work, we tackle model efficiency by exploiting redundancy in the \textit{implicit structure} of the building blocks of convolutional neural networks. We start our analysis by introducing a general definition of Composite Kernel structures that enable the execution of convolution operations in the form of efficient, scaled, sum-pooling components. As its special case, we propose \textit{Structured Convolutions} and show that these allow decomposition of the convolution operation into a sum-pooling operation followed by a convolution with significantly lower complexity and fewer weights. We show how this decomposition can be applied to 2D and 3D kernels as well as the fully-connected layers. Furthermore, we present a Structural Regularization loss that promotes neural network layers to leverage on this desired structure in a way that, after training, they can be decomposed with negligible performance loss. By applying our method to a wide range of CNN architectures, we demonstrate "structured" versions of the ResNets that are up to 2$\times$ smaller and a new Structured-MobileNetV2 that is more efficient while staying within an accuracy loss of 1% on ImageNet and CIFAR-10 datasets. We also show similar structured versions of EfficientNet on ImageNet and HRNet architecture for semantic segmentation on the Cityscapes dataset. Our method performs equally well or superior in terms of the complexity reduction in comparison to the existing tensor decomposition and channel pruning methods.

Yash Bhalgat, Jinwon Lee, Markus Nagel et al.

Unlike ReLU, newer activation functions (like Swish, H-swish, Mish) that are frequently employed in popular efficient architectures can also result in negative activation values, with skewed positive and negative ranges. Typical learnable quantization schemes [PACT, LSQ] assume unsigned quantization for activations and quantize all negative activations to zero which leads to significant loss in performance. Naively using signed quantization to accommodate these negative values requires an extra sign bit which is expensive for low-bit (2-, 3-, 4-bit) quantization. To solve this problem, we propose LSQ+, a natural extension of LSQ, wherein we introduce a general asymmetric quantization scheme with trainable scale and offset parameters that can learn to accommodate the negative activations. Gradient-based learnable quantization schemes also commonly suffer from high instability or variance in the final training performance, hence requiring a great deal of hyper-parameter tuning to reach a satisfactory performance. LSQ+ alleviates this problem by using an MSE-based initialization scheme for the quantization parameters. We show that this initialization leads to significantly lower variance in final performance across multiple training runs. Overall, LSQ+ shows state-of-the-art results for EfficientNet and MixNet and also significantly outperforms LSQ for low-bit quantization of neural nets with Swish activations (e.g.: 1.8% gain with W4A4 quantization and upto 5.6% gain with W2A2 quantization of EfficientNet-B0 on ImageNet dataset). To the best of our knowledge, ours is the first work to quantize such architectures to extremely low bit-widths.

Kambiz Azarian, Yash Bhalgat, Jinwon Lee et al.

This paper presents a novel differentiable method for unstructured weight pruning of deep neural networks. Our learned-threshold pruning (LTP) method learns per-layer thresholds via gradient descent, unlike conventional methods where they are set as input. Making thresholds trainable also makes LTP computationally efficient, hence scalable to deeper networks. For example, it takes $30$ epochs for LTP to prune ResNet50 on ImageNet by a factor of $9.1$. This is in contrast to other methods that search for per-layer thresholds via a computationally intensive iterative pruning and fine-tuning process. Additionally, with a novel differentiable $L_0$ regularization, LTP is able to operate effectively on architectures with batch-normalization. This is important since $L_1$ and $L_2$ penalties lose their regularizing effect in networks with batch-normalization. Finally, LTP generates a trail of progressively sparser networks from which the desired pruned network can be picked based on sparsity and performance requirements. These features allow LTP to achieve competitive compression rates on ImageNet networks such as AlexNet ($26.4\times$ compression with $79.1\%$ Top-5 accuracy) and ResNet50 ($9.1\times$ compression with $92.0\%$ Top-5 accuracy). We also show that LTP effectively prunes modern \textit{compact} architectures, such as EfficientNet, MobileNetV2 and MixNet.

Jangho Kim, Yash Bhalgat, Jinwon Lee et al.

Quantization and Knowledge distillation (KD) methods are widely used to reduce memory and power consumption of deep neural networks (DNNs), especially for resource-constrained edge devices. Although their combination is quite promising to meet these requirements, it may not work as desired. It is mainly because the regularization effect of KD further diminishes the already reduced representation power of a quantized model. To address this short-coming, we propose Quantization-aware Knowledge Distillation (QKD) wherein quantization and KD are care-fully coordinated in three phases. First, Self-studying (SS) phase fine-tunes a quantized low-precision student network without KD to obtain a good initialization. Second, Co-studying (CS) phase tries to train a teacher to make it more quantizaion-friendly and powerful than a fixed teacher. Finally, Tutoring (TU) phase transfers knowledge from the trained teacher to the student. We extensively evaluate our method on ImageNet and CIFAR-10/100 datasets and show an ablation study on networks with both standard and depthwise-separable convolutions. The proposed QKD outperformed existing state-of-the-art methods (e.g., 1.3% improvement on ResNet-18 with W4A4, 2.6% on MobileNetV2 with W4A4). Additionally, QKD could recover the full-precision accuracy at as low as W3A3 quantization on ResNet and W6A6 quantization on MobilenetV2.

Yash Bhalgat, Zhe Liu, Pritam Gundecha et al.

Given that labeled data is expensive to obtain in real-world scenarios, many semi-supervised algorithms have explored the task of exploitation of unlabeled data. Traditional tri-training algorithm and tri-training with disagreement have shown promise in tasks where labeled data is limited. In this work, we introduce a new paradigm for tri-training, mimicking the real world teacher-student learning process. We show that the adaptive teacher-student thresholds used in the proposed method provide more control over the learning process with higher label quality. We perform evaluation on SemEval sentiment analysis task and provide comprehensive comparisons over experimental settings containing varied labeled versus unlabeled data rates. Experimental results show that our method outperforms other strong semi-supervised baselines, while requiring less number of labeled training samples.

Yash Bhalgat, Meet Shah, Suyash Awate

For medical image segmentation, most fully convolutional networks (FCNs) need strong supervision through a large sample of high-quality dense segmentations, which is taxing in terms of costs, time and logistics involved. This burden of annotation can be alleviated by exploiting weak inexpensive annotations such as bounding boxes and anatomical landmarks. However, it is very difficult to \textit{a priori} estimate the optimal balance between the number of annotations needed for each supervision type that leads to maximum performance with the least annotation cost. To optimize this cost-performance trade off, we present a budget-based cost-minimization framework in a mixed-supervision setting via dense segmentations, bounding boxes, and landmarks. We propose a linear programming (LP) formulation combined with uncertainty and similarity based ranking strategy to judiciously select samples to be annotated next for optimal performance. In the results section, we show that our proposed method achieves comparable performance to state-of-the-art approaches with significantly reduced cost of annotations.

Yash Bhalgat

This paper describes two of my best performing approaches on the Content-based Video Relevance Prediction challenge. In the FusedLSTM based approach, the inception-pool3 and the C3D-pool5 features are combined using an LSTM and a dense layer to form embeddings with the objective to minimize the triplet loss function. In the second approach, an Online Kernel Similarity Learning method is proposed to learn a non-linear similarity measure to adhere the relevance training data. The last section gives a complete comparison of all the approaches implemented during this challenge, including the one presented in the baseline paper.

Yash Bhalgat, Mandar Kulkarni, Shirish Karande et al.

Document digitization is becoming increasingly crucial. In this work, we propose a shape based approach for automatic stamp verification/detection in document images using an unsupervised feature learning. Given a small set of training images, our algorithm learns an appropriate shape representation using an unsupervised clustering. Experimental results demonstrate the effectiveness of our framework in challenging scenarios.