Image similarity has been extensively studied in computer vision. In recently years, machine-learned models have shown their ability to encode more semantics than traditional multivariate metrics. However, in labelling similarity, assigning a numerical score to a pair of images is less intuitive than determining if an image A is closer to a reference image R than another image B. In this work, we present a novel approach for building an image similarity model based on labelled data in the form of A:R vs B:R. We address the challenges of sparse sampling in the image space (R, A, B) and biases in the models trained with context-based data by using an ensemble model. In particular, we employed two ML techniques to construct such an ensemble model, namely dimensionality reduction and MLP regressors. Our testing results show that the ensemble model constructed performs ~5% better than the best individual context-sensitive models. They also performed better than the model trained with mixed imagery data as well as existing similarity models, e.g., CLIP and DINO. This work demonstrate that context-based labelling and model training can be effective when an appropriate ensemble approach is used to alleviate the limitation due to sparse sampling.
AI-generated medical images are gaining growing popularity due to their potential to address the data scarcity challenge in the real world. However, the issue of accurate identification of these synthetic images, particularly when they exhibit remarkable realism with their real copies, remains a concern. To mitigate this challenge, image generators such as DALLE and Imagen, have integrated digital watermarks aimed at facilitating the discernment of synthetic images' authenticity. These watermarks are embedded within the image pixels and are invisible to the human eye while remains their detectability. Nevertheless, a comprehensive investigation into the potential impact of these invisible watermarks on the utility of synthetic medical images has been lacking. In this study, we propose the incorporation of invisible watermarks into synthetic medical images and seek to evaluate their efficacy in the context of downstream classification tasks. Our goal is to pave the way for discussions on the viability of such watermarks in boosting the detectability of synthetic medical images, fortifying ethical standards, and safeguarding against data pollution and potential scams.
This research paper presents a unique Bengali OCR system with some capabilities. The system excels in reconstructing document layouts while preserving structure, alignment, and images. It incorporates advanced image and signature detection for accurate extraction. Specialized models for word segmentation cater to diverse document types, including computer-composed, letterpress, typewriter, and handwritten documents. The system handles static and dynamic handwritten inputs, recognizing various writing styles. Furthermore, it has the ability to recognize compound characters in Bengali. Extensive data collection efforts provide a diverse corpus, while advanced technical components optimize character and word recognition. Additional contributions include image, logo, signature and table recognition, perspective correction, layout reconstruction, and a queuing module for efficient and scalable processing. The system demonstrates outstanding performance in efficient and accurate text extraction and analysis.
In light of recent advances in multimodal Large Language Models (LLMs), there is increasing attention to scaling them from image-text data to more informative real-world videos. Compared to static images, video poses unique challenges for effective large-scale pre-training due to the modeling of its spatiotemporal dynamics. In this paper, we address such limitations in video-language pre-training with an efficient video decomposition that represents each video as keyframes and temporal motions. These are then adapted to an LLM using well-designed tokenizers that discretize visual and temporal information as a few tokens, thus enabling unified generative pre-training of videos, images, and text. At inference, the generated tokens from the LLM are carefully recovered to the original continuous pixel space to create various video content. Our proposed framework is both capable of comprehending and generating image and video content, as demonstrated by its competitive performance across 13 multimodal benchmarks in image and video understanding and generation. Our code and models will be available at https://video-lavit.github.io.
Scaling up contrastive language-image pretraining (CLIP) is critical for empowering both vision and multimodal models. We present EVA-CLIP-18B, the largest and most powerful open-source CLIP model to date, with 18-billion parameters. With only 6-billion training samples seen, EVA-CLIP-18B achieves an exceptional 80.7% zero-shot top-1 accuracy averaged across 27 widely recognized image classification benchmarks, outperforming its forerunner EVA-CLIP (5-billion parameters) and other open-source CLIP models by a large margin. Remarkably, we observe a consistent performance improvement with the model size scaling of EVA-CLIP, despite maintaining a constant training dataset of 2-billion image-text pairs from LAION-2B and COYO-700M. This dataset is openly available and much smaller than the in-house datasets (e.g., DFN-5B, WebLI-10B) employed in other state-of-the-art CLIP models. EVA-CLIP-18B demonstrates the potential of EVA-style weak-to-strong visual model scaling. With our model weights made publicly available, we hope to facilitate future research in vision and multimodal foundation models.
Recently, the cross-modal pretraining model has been employed to produce meaningful pseudo-labels to supervise the training of an image clustering model. However, numerous erroneous alignments in a cross-modal pre-training model could produce poor-quality pseudo-labels and degrade clustering performance. To solve the aforementioned issue, we propose a novel \textbf{Multi-level Cross-modal Alignment} method to improve the alignments in a cross-modal pretraining model for downstream tasks, by building a smaller but better semantic space and aligning the images and texts in three levels, i.e., instance-level, prototype-level, and semantic-level. Theoretical results show that our proposed method converges, and suggests effective means to reduce the expected clustering risk of our method. Experimental results on five benchmark datasets clearly show the superiority of our new method.
Reward finetuning has emerged as a promising approach to aligning foundation models with downstream objectives. Remarkable success has been achieved in the language domain by using reinforcement learning (RL) to maximize rewards that reflect human preference. However, in the vision domain, existing RL-based reward finetuning methods are limited by their instability in large-scale training, rendering them incapable of generalizing to complex, unseen prompts. In this paper, we propose Proximal Reward Difference Prediction (PRDP), enabling stable black-box reward finetuning for diffusion models for the first time on large-scale prompt datasets with over 100K prompts. Our key innovation is the Reward Difference Prediction (RDP) objective that has the same optimal solution as the RL objective while enjoying better training stability. Specifically, the RDP objective is a supervised regression objective that tasks the diffusion model with predicting the reward difference of generated image pairs from their denoising trajectories. We theoretically prove that the diffusion model that obtains perfect reward difference prediction is exactly the maximizer of the RL objective. We further develop an online algorithm with proximal updates to stably optimize the RDP objective. In experiments, we demonstrate that PRDP can match the reward maximization ability of well-established RL-based methods in small-scale training. Furthermore, through large-scale training on text prompts from the Human Preference Dataset v2 and the Pick-a-Pic v1 dataset, PRDP achieves superior generation quality on a diverse set of complex, unseen prompts whereas RL-based methods completely fail.
Image animation aims to bring static images to life according to driving videos and create engaging visual content that can be used for various purposes such as animation, entertainment, and education. Recent unsupervised methods utilize affine and thin-plate spline transformations based on keypoints to transfer the motion in driving frames to the source image. However, limited by the expressive power of the transformations used, these methods always produce poor results when the gap between the motion in the driving frame and the source image is large. To address this issue, we propose to model motion from the source image to the driving frame in highly-expressive diffeomorphism spaces. Firstly, we introduce Continuous Piecewise-Affine based (CPAB) transformation to model the motion and present a well-designed inference algorithm to generate CPAB transformation from control keypoints. Secondly, we propose a SAM-guided keypoint semantic loss to further constrain the keypoint extraction process and improve the semantic consistency between the corresponding keypoints on the source and driving images. Finally, we design a structure alignment loss to align the structure-related features extracted from driving and generated images, thus helping the generator generate results that are more consistent with the driving action. Extensive experiments on four datasets demonstrate the effectiveness of our method against state-of-the-art competitors quantitatively and qualitatively. Code will be publicly available at: https://github.com/DevilPG/AAAI2024-CPABMM.
While branching network structures abound in nature, their objective analysis is more difficult than expected because existing quantitative methods often rely on the subjective judgment of branch structures. This problem is particularly pronounced when dealing with images comprising discrete particles. Here we propose an objective framework for quantitative analysis of branching networks by introducing the mathematical definitions for internal and external structures based on topological data analysis, specifically, persistent homology. We compare persistence diagrams constructed from images with and without plots on the convex hull. The unchanged points in the two diagrams are the internal structures and the difference between the two diagrams is the external structures. We construct a mathematical theory for our method and show that the internal structures have a monotonicity relationship with respect to the plots on the convex hull, while the external structures do not. This is the phenomenon related to the resolution of the image. Our method can be applied to a wide range of branch structures in biology, enabling objective analysis of numbers, spatial distributions, sizes, and more. Additionally, our method has the potential to be combined with other tools in topological data analysis, such as the generalized persistence landscape.
Multimodal Large Language Models (MLLMs) have recently shown remarkable perceptual capability in answering visual questions, however, little is known about the limits of their perception. In particular, while prior works have provided anecdotal evidence of MLLMs' sensitivity to object size, this phenomenon and its underlying causes have not been explored comprehensively. In this work, we quantitatively study the perception of small visual objects in several state-of-the-art MLLMs and reveal a pervasive limitation in answering questions about small objects in images. Next, we identify four independent factors that can contribute to this limitation -- object quality, size, distractors, and location -- and conduct controlled intervention studies to measure the effect of each factor on MLLMs' perception. In particular, we find that lower object quality and smaller object size can both independently reduce MLLMs' ability to answer visual questions. More surprisingly, we find that the location of the object in the image and the presence of visual distractors can also significantly reduce MLLMs' question answering accuracy. Our study provides a better understanding of the perceptual limitation of MLLMs and contributes new evaluation protocols for analyzing the perception of future MLLMs. To facilitate further investigations, we release our code and data.