Inna Stainvas

Alon Kaya, Igal Bilik, Inna Stainvas

Vision-transformers (ViTs) and large-scale convolution-neural-networks (CNNs) have reshaped computer vision through pretrained feature representations that enable strong transfer learning for diverse tasks. However, their efficiency as backbone architectures for geometric estimation tasks involving image deformations in low-data regimes remains an open question. This work considers two such tasks: 1) estimating 2D rigid transformations between pairs of images and 2) predicting the fundamental matrix for stereo image pairs, an important problem in various applications, such as autonomous mobility, robotics, and 3D scene reconstruction. Addressing this intriguing question, this work systematically compares large-scale CNNs (ResNet, EfficientNet, CLIP-ResNet) with ViT-based foundation models (CLIP-ViT variants and DINO) in various data size settings, including few-shot scenarios. These pretrained models are optimized for classification or contrastive learning, encouraging them to focus mostly on high-level semantics. The considered tasks require balancing local and global features differently, challenging the straightforward adoption of these models as the backbone. Empirical comparative analysis shows that, similar to training from scratch, ViTs outperform CNNs during refinement in large downstream-data scenarios. However, in small data scenarios, the inductive bias and smaller capacity of CNNs improve their performance, allowing them to match that of a ViT. Moreover, ViTs exhibit stronger generalization in cross-domain evaluation where the data distribution changes. These results emphasize the importance of carefully selecting model architectures for refinement, motivating future research towards hybrid architectures that balance local and global representations.

Eliraz Orfaig, Inna Stainvas, Igal Bilik

This work introduces RGBX-DiffusionDet, an object detection framework extending the DiffusionDet model to fuse the heterogeneous 2D data (X) with RGB imagery via an adaptive multimodal encoder. To enable cross-modal interaction, we design the dynamic channel reduction within a convolutional block attention module (DCR-CBAM), which facilitates cross-talk between subnetworks by dynamically highlighting salient channel features. Furthermore, the dynamic multi-level aggregation block (DMLAB) is proposed to refine spatial feature representations through adaptive multiscale fusion. Finally, novel regularization losses that enforce channel saliency and spatial selectivity are introduced, leading to compact and discriminative feature embeddings. Extensive experiments using RGB-Depth (KITTI), a novel annotated RGB-Polarimetric dataset, and RGB-Infrared (M$^3$FD) benchmark dataset were conducted. We demonstrate consistent superiority of the proposed approach over the baseline RGB-only DiffusionDet. The modular architecture maintains the original decoding complexity, ensuring efficiency. These results establish the proposed RGBX-DiffusionDet as a flexible multimodal object detection approach, providing new insights into integrating diverse 2D sensing modalities into diffusion-based detection pipelines.

Eliraz Orfaig, Inna Stainvas, Igal Bilik

Vision-based autonomous driving requires reliable and efficient object detection. This work proposes a DiffusionDet-based framework that exploits data fusion from the monocular camera and depth sensor to provide the RGB and depth (RGB-D) data. Within this framework, ground truth bounding boxes are randomly reshaped as part of the training phase, allowing the model to learn the reverse diffusion process of noise addition. The system methodically enhances a randomly generated set of boxes at the inference stage, guiding them toward accurate final detections. By integrating the textural and color features from RGB images with the spatial depth information from the LiDAR sensors, the proposed framework employs a feature fusion that substantially enhances object detection of automotive targets. The $2.3$ AP gain in detecting automotive targets is achieved through comprehensive experiments using the KITTI dataset. Specifically, the improved performance of the proposed approach in detecting small objects is demonstrated.

Inna Stainvas, Alexandra Manevitch, Isaac Leichter

This paper deals with the multiple annotation problem in medical application of cancer detection in digital images. The main assumption is that though images are labeled by many experts, the number of images read by the same expert is not large. Thus differing with the existing work on modeling each expert and ground truth simultaneously, the multi annotation information is used in a soft manner. The multiple labels from different experts are used to estimate the probability of the findings to be marked as malignant. The learning algorithm minimizes the Kullback Leibler (KL) divergence between the modeled probabilities and desired ones constraining the model to be compact. The probabilities are modeled by logit regression and multiple instance learning concept is used by us. Experiments on a real-life computer aided diagnosis (CAD) problem for CXR CAD lung cancer detection demonstrate that the proposed algorithm leads to similar results as learning with a binary RVMMIL classifier or a mixture of binary RVMMIL models per annotator. However, this model achieves a smaller complexity and is more preferable in practice.