Computer vision

Computer vision traditionally concentrates on acknowledging universally agreed concepts like animals, vehicles, or specific objects. However, real-world applications often need to identify variable subjective concepts like predicting emotions, determining aesthetic appeal, or regulating content. What is considered "unsafe" content or "gourmet" food differs greatly among individuals, hence the increasing demand for user-centric training frameworks that…

The evolution of Multimodal Large Language Models (MLLMs) has been significant, particularly those models that blend language and vision modalities (LVMs). There has been growing interest in applying MLLMs in various fields like computer vision tasks and integrating them into complex pipelines. Despite some models like ShareGPTV performing well in data annotation tasks, their practical…

Computer vision researchers frequently concentrate on developing powerful encoder networks for self-supervised learning (SSL) methods, intending to generate image representations. However, the predictive part of the model, which potentially contains valuable information, is often overlooked post-pretraining. This research introduces a distinctive approach that repurposes the predictive model for various downstream vision tasks rather than discarding…

Recent advancements in large vision-language models (VLMs) have demonstrated great potential in performing multimodal tasks. However, these models have shortcomings when it comes to fine-grained region grounding, inter-object spatial relations, and compositional reasoning. These limitations affect the model's capability to follow visual prompts like bounding boxes that spotlight vital regions. Challenged by these limitations, researchers at…

Neural Architecture Search (NAS) is a process that utilizes machine learning to automate the design of neural networks. This development has marked a significant shift from traditional manual design processes and is considered pivotal in paving the way for future advancements in autonomous machine learning. Despite these benefits, adopting NAS in the past has been…

Deep Neural Networks (DNNs) have demonstrated substantial prowess in improving surgical precision by accurately identifying robotic instruments and tissues through semantic segmentation. However, DNNs grapple with catastrophic forgetting, signifying a rapid performance decline on previously learned tasks when new ones are introduced. This poses significant problems, especially in cases where old data is not accessible…

Large Vision-Language Models (LVLMs), which combine powerful language and vision encoders, have shown excellent proficiency in tasks involving real-world images. However, they have generally struggled with abstract ideas, primarily due to their lack of exposure to domain-specific data during training. This is particularly true for areas requiring abstract reasoning, such as physics and mathematics. To address…

In recent years, large language models such as LLaMA, largely based on transformer architectures, have significantly influenced the field of natural language processing. This raises the question of whether the transformer architecture can be applied effectively to process 2D images. In response, a paper introduces VisionLLaMA, a vision transformer that seeks to bridge language and…