Nighttime camera-based depth estimation is a highly challenging task, especially for autonomous driving applications, where accurate depth perception is essential for ensuring safe navigation. We aim to improve the reliability of perception systems at night time, where models trained on daytime data often fail in the absence of precise but costly LiDAR sensors. In this work, we introduce Light Enhanced Depth (LED), a novel cost-effective approach that significantly improves depth estimation in low-light environments by harnessing a pattern projected by high definition headlights available in modern vehicles. LED leads to significant performance boosts across multiple depth-estimation architectures (encoder-decoder, Adabins, DepthFormer) both on synthetic and real datasets. Furthermore, increased performances beyond illuminated areas reveal a holistic enhancement in scene understanding. Finally, we release the Nighttime Synthetic Drive Dataset, a new synthetic and photo-realistic nighttime dataset, which comprises 49,990 comprehensively annotated images.
RGB-D has gradually become a crucial data source for understanding complex scenes in assisted driving. However, existing studies have paid insufficient attention to the intrinsic spatial properties of depth maps. This oversight significantly impacts the attention representation, leading to prediction errors caused by attention shift issues. To this end, we propose a novel learnable Depth interaction Pyramid Transformer (DiPFormer) to explore the effectiveness of depth. Firstly, we introduce Depth Spatial-Aware Optimization (Depth SAO) as offset to represent real-world spatial relationships. Secondly, the similarity in the feature space of RGB-D is learned by Depth Linear Cross-Attention (Depth LCA) to clarify spatial differences at the pixel level. Finally, an MLP Decoder is utilized to effectively fuse multi-scale features for meeting real-time requirements. Comprehensive experiments demonstrate that the proposed DiPFormer significantly addresses the issue of attention misalignment in both road detection (+7.5%) and semantic segmentation (+4.9% / +1.5%) tasks. DiPFormer achieves state-of-the-art performance on the KITTI (97.57% F-score on KITTI road and 68.74% mIoU on KITTI-360) and Cityscapes (83.4% mIoU) datasets.
Neural networks achieve state-of-the-art performance in many supervised learning tasks when the training data distribution matches the test data distribution. However, their performance drops significantly under domain (covariate) shift, a prevalent issue in medical image segmentation due to varying acquisition settings across different scanner models and protocols. Recently, foundational models (FMs) trained on large datasets have gained attention for their ability to be adapted for downstream tasks and achieve state-of-the-art performance with excellent generalization capabilities on natural images. However, their effectiveness in medical image segmentation remains underexplored. In this paper, we investigate the domain generalization performance of various FMs, including DinoV2, SAM, MedSAM, and MAE, when fine-tuned using various parameter-efficient fine-tuning (PEFT) techniques such as Ladder and Rein (+LoRA) and decoder heads. We introduce a novel decode head architecture, HQHSAM, which simply integrates elements from two state-of-the-art decoder heads, HSAM and HQSAM, to enhance segmentation performance. Our extensive experiments on multiple datasets, encompassing various anatomies and modalities, reveal that FMs, particularly with the HQHSAM decode head, improve domain generalization for medical image segmentation. Moreover, we found that the effectiveness of PEFT techniques varies across different FMs. These findings underscore the potential of FMs to enhance the domain generalization performance of neural networks in medical image segmentation across diverse clinical settings, providing a solid foundation for future research. Code and models are available for research purposes at \url{https://github.com/kerem-cekmeceli/Foundation-Models-for-Medical-Imagery}.
We propose a new method for generating realistic datasets with distribution shifts using any decoder-based generative model. Our approach systematically creates datasets with varying intensities of distribution shifts, facilitating a comprehensive analysis of model performance degradation. We then use these generated datasets to evaluate the performance of various commonly used networks and observe a consistent decline in performance with increasing shift intensity, even when the effect is almost perceptually unnoticeable to the human eye. We see this degradation even when using data augmentations. We also find that enlarging the training dataset beyond a certain point has no effect on the robustness and that stronger inductive biases increase robustness.
Initially designed for image generation, diffusion models can also be effectively applied to various tasks, including semantic segmentation. However, most existing diffusion-based approaches for semantic segmentation operate in high-dimensional pixel space, demanding a lot of computing and memory resources during training and inference. This paper makes the first attempt to utilize latent diffusion models for semantic segmentation. Specifically, we propose a fast yet effective latent diffusion model and valuate it on medical image segmentation tasks. Firstly, we train a Variational Autoencoder (VAE) network to convert binary image masks into compact latent vectors. The diffusion process can then be executed in this low-dimensional latent space and thus drastically accelerated. Subsequently, we employ the VAE decoder to reconstruct a precise prediction map from the latent output vector produced by the diffusion process. Eventually, we refine the final segmentation results through a straightforward post-processing step using morphological operations. We report our results on two public datasets, including colon polyp images and skin cancer images. Experiments show that our approach achieves competitive accuracy compared to traditional diffusion models while having much better training and inference speed, as well as much more efficient memory consumption.
Bimanual manipulation presents unique challenges compared to unimanual tasks due to the complexity of coordinating two robotic arms. In this paper, we introduce InterACT: Inter-dependency aware Action Chunking with Hierarchical Attention Transformers, a novel imitation learning framework designed specifically for bimanual manipulation. InterACT leverages hierarchical attention mechanisms to effectively capture inter-dependencies between dual-arm joint states and visual inputs. The framework comprises a Hierarchical Attention Encoder, which processes multi-modal inputs through segment-wise and cross-segment attention mechanisms, and a Multi-arm Decoder that generates each arm's action predictions in parallel, while sharing information between the arms through synchronization blocks by providing the other arm's intermediate output as context. Our experiments, conducted on various simulated and real-world bimanual manipulation tasks, demonstrate that InterACT outperforms existing methods. Detailed ablation studies further validate the significance of key components, including the impact of CLS tokens, cross-segment encoders, and synchronization blocks on task performance. We provide supplementary materials and videos on our project page.
Medical image segmentation, a crucial task in computer vision, facilitates the automated delineation of anatomical structures and pathologies, supporting clinicians in diagnosis, treatment planning, and disease monitoring. Notably, transformers employing shifted window-based self-attention have demonstrated exceptional performance. However, their reliance on local window attention limits the fusion of local and global contextual information, crucial for segmenting microtumors and miniature organs. To address this limitation, we propose the Adaptive Semantic Segmentation Network (ASSNet), a transformer architecture that effectively integrates local and global features for precise medical image segmentation. ASSNet comprises a transformer-based U-shaped encoder-decoder network. The encoder utilizes shifted window self-attention across five resolutions to extract multi-scale features, which are then propagated to the decoder through skip connections. We introduce an augmented multi-layer perceptron within the encoder to explicitly model long-range dependencies during feature extraction. Recognizing the constraints of conventional symmetrical encoder-decoder designs, we propose an Adaptive Feature Fusion (AFF) decoder to complement our encoder. This decoder incorporates three key components: the Long Range Dependencies (LRD) block, the Multi-Scale Feature Fusion (MFF) block, and the Adaptive Semantic Center (ASC) block. These components synergistically facilitate the effective fusion of multi-scale features extracted by the decoder while capturing long-range dependencies and refining object boundaries. Comprehensive experiments on diverse medical image segmentation tasks, including multi-organ, liver tumor, and bladder tumor segmentation, demonstrate that ASSNet achieves state-of-the-art results. Code and models are available at: \url{https://github.com/lzeeorno/ASSNet}.
Automated pavement monitoring using computer vision can analyze pavement conditions more efficiently and accurately than manual methods. Accurate segmentation is essential for quantifying the severity and extent of pavement defects and consequently, the overall condition index used for prioritizing rehabilitation and maintenance activities. Deep learning-based segmentation models are however, often supervised and require pixel-level annotations, which can be costly and time-consuming. While the recent evolution of zero-shot segmentation models can generate pixel-wise labels for unseen classes without any training data, they struggle with irregularities of cracks and textured pavement backgrounds. This research proposes a zero-shot segmentation model, PaveSAM, that can segment pavement distresses using bounding box prompts. By retraining SAM's mask decoder with just 180 images, pavement distress segmentation is revolutionized, enabling efficient distress segmentation using bounding box prompts, a capability not found in current segmentation models. This not only drastically reduces labeling efforts and costs but also showcases our model's high performance with minimal input, establishing the pioneering use of SAM in pavement distress segmentation. Furthermore, researchers can use existing open-source pavement distress images annotated with bounding boxes to create segmentation masks, which increases the availability and diversity of segmentation pavement distress datasets.
We propose a scaling law hypothesis for multimodal models processing text, audio, images, and video within a shared token and embedding space. Our framework predicts model performance based on modality-specific compression and tokenization efficiency, extending established scaling laws from text-based decoder models to mixed-modality systems. We explore whether leveraging more training data in multiple modalities can reduce the size of the multimodal model, enabling efficient deployment on resource-constrained devices.
In the realm of Large Language Models (LLMs), the ability to process long contexts is increasingly crucial for tasks such as multi-round dialogues, code generation, and document summarization. This paper addresses the challenges of enhancing the long-context performance, reducing computational complexity, and leveraging pretrained models collectively termed the "impossible triangle." We introduce E2LLM (Encoder Elongated Large Language Models), a novel approach that effectively navigates this paradox. The method involves splitting long contexts into chunks, compressing each into embedding vectors via a pretrained text encoder, and utilizing an adapter to align these representations with a decoder-only LLM. Two training objectives, focusing on reconstruction of the encoder output and long-context instruction fine-tuning, are employed to facilitate the understanding of soft prompts by the LLM. Experimental results demonstrate that E2LLM achieves superior performance in long-context scenarios while balancing efficiency, performance, and compatibility with pretrained models. Our framework thus represents a significant advancement in the field, contributing to effective long-text modeling.
Recently, the 3D Gaussian Splatting (3D-GS) method has achieved great success in novel view synthesis, providing real-time rendering while ensuring high-quality rendering results. However, this method faces challenges in modeling specular reflections and handling anisotropic appearance components, especially in dealing with view-dependent color under complex lighting conditions. Additionally, 3D-GS uses spherical harmonic to learn the color representation, which has limited ability to represent complex scenes. To overcome these challenges, we introduce Lantent-SpecGS, an approach that utilizes a universal latent neural descriptor within each 3D Gaussian. This enables a more effective representation of 3D feature fields, including appearance and geometry. Moreover, two parallel CNNs are designed to decoder the splatting feature maps into diffuse color and specular color separately. A mask that depends on the viewpoint is learned to merge these two colors, resulting in the final rendered image. Experimental results demonstrate that our method obtains competitive performance in novel view synthesis and extends the ability of 3D-GS to handle intricate scenarios with specular reflections.
Existing few-shot segmentation (FSS) methods mainly focus on designing novel support-query matching and self-matching mechanisms to exploit implicit knowledge in pre-trained backbones. However, the performance of these methods is often constrained by models pre-trained on classification tasks. The exploration of what types of pre-trained models can provide more beneficial implicit knowledge for FSS remains limited. In this paper, inspired by the representation consistency of foundational computer vision models, we develop a FSS framework based on foundation models. To be specific, we propose a simple approach to extract implicit knowledge from foundation models to construct coarse correspondence and introduce a lightweight decoder to refine coarse correspondence for fine-grained segmentation. We systematically summarize the performance of various foundation models on FSS and discover that the implicit knowledge within some of these models is more beneficial for FSS than models pre-trained on classification tasks. Extensive experiments on two widely used datasets demonstrate the effectiveness of our approach in leveraging the implicit knowledge of foundation models. Notably, the combination of DINOv2 and DFN exceeds previous state-of-the-art methods by 17.5% on COCO-20i. Code is available at https://github.com/DUT-CSJ/FoundationFSS.
Unbiased representation learning is still an object of study under specific applications and contexts. Novel architectures are usually crafted to resolve particular problems using mixtures of fundamental pieces. This paper presents different image feature extraction mechanisms that work together with residual connections to encode perceptual image information in an autoencoder configuration. We use image data that aims to support a larger research agenda dealing with issues regarding criminal activity in consumer-to-consumer online platforms. Preliminary results suggest that the proposed architecture can learn rich spaces using ours and other image datasets resolving important challenges that are identified.
Generative recommendation systems have gained increasing attention as an innovative approach that directly generates item identifiers for recommendation tasks. Despite their potential, a major challenge is the effective construction of item identifiers that align well with recommender systems. Current approaches often treat item tokenization and generative recommendation training as separate processes, which can lead to suboptimal performance. To overcome this issue, we introduce ETEGRec, a novel End-To-End Generative Recommender that unifies item tokenization and generative recommendation into a cohesive framework. Built on a dual encoder-decoder architecture, ETEGRec consists of an item tokenizer and a generative recommender. To enable synergistic interaction between these components, we propose a recommendation-oriented alignment strategy, which includes two key optimization objectives: sequence-item alignment and preference-semantic alignment. These objectives tightly couple the learning processes of the item tokenizer and the generative recommender, fostering mutual enhancement. Additionally, we develop an alternating optimization technique to ensure stable and efficient end-to-end training of the entire framework. Extensive experiments demonstrate the superior performance of our approach compared to traditional sequential recommendation models and existing generative recommendation baselines.
Predictive learning for spatio-temporal processes (PL-STP) on complex spatial domains plays a critical role in various scientific and engineering fields, with its essence being the construction of operators between infinite-dimensional function spaces. This paper focuses on the unequal-domain mappings in PL-STP and categorising them into increase-domain and decrease-domain mapping. Recent advances in deep learning have revealed the great potential of neural operators (NOs) to learn operators directly from observational data. However, existing NOs require input space and output space to be the same domain, which pose challenges in ensuring predictive accuracy and stability for unequal-domain mappings. To this end, this study presents a general reduced-order neural operator named Reduced-Order Neural Operator on Riemannian Manifolds (RO-NORM), which consists of two parts: the unequal-domain encoder/decoder and the same-domain approximator. Motivated by the variable separation in classical modal decomposition, the unequal-domain encoder/decoder uses the pre-computed bases to reformulate the spatio-temporal function as a sum of products between spatial (or temporal) bases and corresponding temporally (or spatially) distributed weight functions, thus the original unequal-domain mapping can be converted into a same-domain mapping. Consequently, the same-domain approximator NORM is applied to model the transformed mapping. The performance of our proposed method has been evaluated on six benchmark cases, including parametric PDEs, engineering and biomedical applications, and compared with four baseline algorithms: DeepONet, POD-DeepONet, PCA-Net, and vanilla NORM. The experimental results demonstrate the superiority of RO-NORM in prediction accuracy and training efficiency for PL-STP.
Nowadays, pre-trained encoders are widely used in medical image segmentation because of their ability to capture complex feature representations. However, the existing models fail to effectively utilize the rich features obtained by the pre-trained encoder, resulting in suboptimal segmentation results. In this work, a novel U-shaped model, called FIF-UNet, is proposed to address the above issue, including three plug-and-play modules. A channel spatial interaction module (CSI) is proposed to obtain informative features by establishing the interaction between encoder stages and corresponding decoder stages. A cascaded conv-SE module (CoSE) is designed to enhance the representation of critical features by adaptively assigning importance weights on different feature channels. A multi-level fusion module (MLF) is proposed to fuse the multi-scale features from the decoder stages, ensuring accurate and robust final segmentation. Comprehensive experiments on the Synapse and ACDC datasets demonstrate that the proposed FIF-UNet outperforms existing state-of-the-art methods, which achieves the highest average DICE of 86.05% and 92.58%, respectively.
Scribble supervised salient object detection (SSSOD) constructs segmentation ability of attractive objects from surroundings under the supervision of sparse scribble labels. For the better segmentation, depth and thermal infrared modalities serve as the supplement to RGB images in the complex scenes. Existing methods specifically design various feature extraction and multi-modal fusion strategies for RGB, RGB-Depth, RGB-Thermal, and Visual-Depth-Thermal image input respectively, leading to similar model flood. As the recently proposed Segment Anything Model (SAM) possesses extraordinary segmentation and prompt interactive capability, we propose an SSSOD family based on SAM, named SSFam, for the combination input with different modalities. Firstly, different modal-aware modulators are designed to attain modal-specific knowledge which cooperates with modal-agnostic information extracted from the frozen SAM encoder for the better feature ensemble. Secondly, a siamese decoder is tailored to bridge the gap between the training with scribble prompt and the testing with no prompt for the stronger decoding ability. Our model demonstrates the remarkable performance among combinations of different modalities and refreshes the highest level of scribble supervised methods and comes close to the ones of fully supervised methods. https://github.com/liuzywen/SSFam
The quality and capabilities of large language models cannot be currently fully assessed with automated, benchmark evaluations. Instead, human evaluations that expand on traditional qualitative techniques from natural language generation literature are required. One recent best-practice consists in using A/B-testing frameworks, which capture preferences of human evaluators for specific models. In this paper we describe a human evaluation experiment focused on the biomedical domain (health, biology, chemistry/pharmacology) carried out at Elsevier. In it a large but not massive (8.8B parameter) decoder-only foundational transformer trained on a relatively small (135B tokens) but highly curated collection of Elsevier datasets is compared to OpenAI's GPT-3.5-turbo and Meta's foundational 7B parameter Llama 2 model against multiple criteria. Results indicate -- even if IRR scores were generally low -- a preference towards GPT-3.5-turbo, and hence towards models that possess conversational abilities, are very large and were trained on very large datasets. But at the same time, indicate that for less massive models training on smaller but well-curated training sets can potentially give rise to viable alternatives in the biomedical domain.
In computer-aided diagnosis tools employed for skin cancer treatment and early diagnosis, skin lesion segmentation is important. However, achieving precise segmentation is challenging due to inherent variations in appearance, contrast, texture, and blurry lesion boundaries. This research presents a robust approach utilizing a dilated convolutional residual network, which incorporates an attention-based spatial feature enhancement block (ASFEB) and employs a guided decoder strategy. In each dilated convolutional residual block, dilated convolution is employed to broaden the receptive field with varying dilation rates. To improve the spatial feature information of the encoder, we employed an attention-based spatial feature enhancement block in the skip connections. The ASFEB in our proposed method combines feature maps obtained from average and maximum-pooling operations. These combined features are then weighted using the active outcome of global average pooling and convolution operations. Additionally, we have incorporated a guided decoder strategy, where each decoder block is optimized using an individual loss function to enhance the feature learning process in the proposed AD-Net. The proposed AD-Net presents a significant benefit by necessitating fewer model parameters compared to its peer methods. This reduction in parameters directly impacts the number of labeled data required for training, facilitating faster convergence during the training process. The effectiveness of the proposed AD-Net was evaluated using four public benchmark datasets. We conducted a Wilcoxon signed-rank test to verify the efficiency of the AD-Net. The outcomes suggest that our method surpasses other cutting-edge methods in performance, even without the implementation of data augmentation strategies.
Facial movements play a crucial role in conveying altitude and intentions, and facial optical flow provides a dynamic and detailed representation of it. However, the scarcity of datasets and a modern baseline hinders the progress in facial optical flow research. This paper proposes FacialFlowNet (FFN), a novel large-scale facial optical flow dataset, and the Decomposed Facial Flow Model (DecFlow), the first method capable of decomposing facial flow. FFN comprises 9,635 identities and 105,970 image pairs, offering unprecedented diversity for detailed facial and head motion analysis. DecFlow features a facial semantic-aware encoder and a decomposed flow decoder, excelling in accurately estimating and decomposing facial flow into head and expression components. Comprehensive experiments demonstrate that FFN significantly enhances the accuracy of facial flow estimation across various optical flow methods, achieving up to an 11% reduction in Endpoint Error (EPE) (from 3.91 to 3.48). Moreover, DecFlow, when coupled with FFN, outperforms existing methods in both synthetic and real-world scenarios, enhancing facial expression analysis. The decomposed expression flow achieves a substantial accuracy improvement of 18% (from 69.1% to 82.1%) in micro-expressions recognition. These contributions represent a significant advancement in facial motion analysis and optical flow estimation. Codes and datasets can be found.