Additive manufacturing (AM) is gaining attention across various industries like healthcare, aerospace, and automotive. However, identifying defects early in the AM process can reduce production costs and improve productivity - a key challenge. This study explored the effectiveness of machine learning (ML) approaches, specifically transfer learning (TL) models, for defect detection in 3D-printed cylinders. Images of cylinders were analyzed using models including VGG16, VGG19, ResNet50, ResNet101, InceptionResNetV2, and MobileNetV2. Performance was compared across two datasets using accuracy, precision, recall, and F1-score metrics. In the first study, VGG16, InceptionResNetV2, and MobileNetV2 achieved perfect scores. In contrast, ResNet50 had the lowest performance, with an average F1-score of 0.32. Similarly, in the second study, MobileNetV2 correctly classified all instances, while ResNet50 struggled with more false positives and fewer true positives, resulting in an F1-score of 0.75. Overall, the findings suggest certain TL models like MobileNetV2 can deliver high accuracy for AM defect classification, although performance varies across algorithms. The results provide insights into model optimization and integration needs for reliable automated defect analysis during 3D printing. By identifying the top-performing TL techniques, this study aims to enhance AM product quality through robust image-based monitoring and inspection.
Diffusion models are powerful generative models that map noise to data using stochastic processes. However, for many applications such as image editing, the model input comes from a distribution that is not random noise. As such, diffusion models must rely on cumbersome methods like guidance or projected sampling to incorporate this information in the generative process. In our work, we propose Denoising Diffusion Bridge Models (DDBMs), a natural alternative to this paradigm based on diffusion bridges, a family of processes that interpolate between two paired distributions given as endpoints. Our method learns the score of the diffusion bridge from data and maps from one endpoint distribution to the other by solving a (stochastic) differential equation based on the learned score. Our method naturally unifies several classes of generative models, such as score-based diffusion models and OT-Flow-Matching, allowing us to adapt existing design and architectural choices to our more general problem. Empirically, we apply DDBMs to challenging image datasets in both pixel and latent space. On standard image translation problems, DDBMs achieve significant improvement over baseline methods, and, when we reduce the problem to image generation by setting the source distribution to random noise, DDBMs achieve comparable FID scores to state-of-the-art methods despite being built for a more general task.
Knowledge-based Visual Question Answering (KB-VQA) requires VQA systems to utilize knowledge from existing knowledge bases to answer visually-grounded questions. Retrieval-Augmented Visual Question Answering (RA-VQA), a strong framework to tackle KB-VQA, first retrieves related documents with Dense Passage Retrieval (DPR) and then uses them to answer questions. This paper proposes Fine-grained Late-interaction Multi-modal Retrieval (FLMR) which significantly improves knowledge retrieval in RA-VQA. FLMR addresses two major limitations in RA-VQA's retriever: (1) the image representations obtained via image-to-text transforms can be incomplete and inaccurate and (2) relevance scores between queries and documents are computed with one-dimensional embeddings, which can be insensitive to finer-grained relevance. FLMR overcomes these limitations by obtaining image representations that complement those from the image-to-text transforms using a vision model aligned with an existing text-based retriever through a simple alignment network. FLMR also encodes images and questions using multi-dimensional embeddings to capture finer-grained relevance between queries and documents. FLMR significantly improves the original RA-VQA retriever's PRRecall@5 by approximately 8\%. Finally, we equipped RA-VQA with two state-of-the-art large multi-modal/language models to achieve $\sim61\%$ VQA score in the OK-VQA dataset.
Previous animatable 3D-aware GANs for human generation have primarily focused on either the human head or full body. However, head-only videos are relatively uncommon in real life, and full body generation typically does not deal with facial expression control and still has challenges in generating high-quality results. Towards applicable video avatars, we present an animatable 3D-aware GAN that generates portrait images with controllable facial expression, head pose, and shoulder movements. It is a generative model trained on unstructured 2D image collections without using 3D or video data. For the new task, we base our method on the generative radiance manifold representation and equip it with learnable facial and head-shoulder deformations. A dual-camera rendering and adversarial learning scheme is proposed to improve the quality of the generated faces, which is critical for portrait images. A pose deformation processing network is developed to generate plausible deformations for challenging regions such as long hair. Experiments show that our method, trained on unstructured 2D images, can generate diverse and high-quality 3D portraits with desired control over different properties.
Recently, there has been significant progress in the development of large models. Following the success of ChatGPT, numerous language models have been introduced, demonstrating remarkable performance. Similar advancements have also been observed in image generation models, such as Google's Imagen model, OpenAI's DALL-E 2, and stable diffusion models, which have exhibited impressive capabilities in generating images. However, similar to large language models, these models still encounter unresolved challenges. Fortunately, the availability of open-source stable diffusion models and their underlying mathematical principles has enabled the academic community to extensively analyze the performance of current image generation models and make improvements based on this stable diffusion framework. This survey aims to examine the existing issues and the current solutions pertaining to image generation models.
Contrastive Language-Image Pre-training (CLIP) is an approach that has advanced research and applications in computer vision, fueling modern recognition systems and generative models. We believe that the main ingredient to the success of CLIP is its data and not the model architecture or pre-training objective. However, CLIP only provides very limited information about its data and how it has been collected, leading to works that aim to reproduce CLIP's data by filtering with its model parameters. In this work, we intend to reveal CLIP's data curation approach and in our pursuit of making it open to the community introduce Metadata-Curated Language-Image Pre-training (MetaCLIP). MetaCLIP takes a raw data pool and metadata (derived from CLIP's concepts) and yields a balanced subset over the metadata distribution. Our experimental study rigorously isolates the model and training settings, concentrating solely on data. MetaCLIP applied to CommonCrawl with 400M image-text data pairs outperforms CLIP's data on multiple standard benchmarks. In zero-shot ImageNet classification, MetaCLIP achieves 70.8% accuracy, surpassing CLIP's 68.3% on ViT-B models. Scaling to 1B data, while maintaining the same training budget, attains 72.4%. Our observations hold across various model sizes, exemplified by ViT-H achieving 80.5%, without any bells-and-whistles. Curation code and training data distribution on metadata is made available at https://github.com/facebookresearch/MetaCLIP.
Neural image classifiers can often learn to make predictions by overly relying on non-predictive features that are spuriously correlated with the class labels in the training data. This leads to poor performance in real-world atypical scenarios where such features are absent. Supplementing the training dataset with images without such spurious features can aid robust learning against spurious correlations via better generalization. This paper presents ASPIRE (Language-guided data Augmentation for SPurIous correlation REmoval), a simple yet effective solution for expanding the training dataset with synthetic images without spurious features. ASPIRE, guided by language, generates these images without requiring any form of additional supervision or existing examples. Precisely, we employ LLMs to first extract foreground and background features from textual descriptions of an image, followed by advanced language-guided image editing to discover the features that are spuriously correlated with the class label. Finally, we personalize a text-to-image generation model to generate diverse in-domain images without spurious features. We demonstrate the effectiveness of ASPIRE on 4 datasets, including the very challenging Hard ImageNet dataset, and 9 baselines and show that ASPIRE improves the classification accuracy of prior methods by 1% - 38%. Code soon at: https://github.com/Sreyan88/ASPIRE.
This work presents CineTransfer, an algorithmic framework that drives a robot to record a video sequence that mimics the cinematographic style of an input video. We propose features that abstract the aesthetic style of the input video, so the robot can transfer this style to a scene with visual details that are significantly different from the input video. The framework builds upon CineMPC, a tool that allows users to control cinematographic features, like subjects' position on the image and the depth of field, by manipulating the intrinsics and extrinsics of a cinematographic camera. However, CineMPC requires a human expert to specify the desired style of the shot (composition, camera motion, zoom, focus, etc). CineTransfer bridges this gap, aiming a fully autonomous cinematographic platform. The user chooses a single input video as a style guide. CineTransfer extracts and optimizes two important style features, the composition of the subject in the image and the scene depth of field, and provides instructions for CineMPC to control the robot to record an output sequence that matches these features as closely as possible. In contrast with other style transfer methods, our approach is a lightweight and portable framework which does not require deep network training or extensive datasets. Experiments with real and simulated videos demonstrate the system's ability to analyze and transfer style between recordings, and are available in the supplementary video.
The lack of ethnic diversity in data has been a limiting factor of face recognition techniques in the literature. This is particularly the case for children where data samples are scarce and presents a challenge when seeking to adapt machine vision algorithms that are trained on adult data to work on children. This work proposes the utilization of image-to-image transformation to synthesize data of different races and thus adjust the ethnicity of children's face data. We consider ethnicity as a style and compare three different Image-to-Image neural network based methods, specifically pix2pix, CycleGAN, and CUT networks to implement Caucasian child data and Asian child data conversion. Experimental validation results on synthetic data demonstrate the feasibility of using image-to-image transformation methods to generate various synthetic child data samples with broader ethnic diversity.
Nowadays, photogrammetrically derived point clouds are widely used in many civilian applications due to their low cost and flexibility in acquisition. Typically, photogrammetric point clouds are assessed through reference data such as LiDAR point clouds. However, when reference data are not available, the assessment of photogrammetric point clouds may be challenging. Since these point clouds are algorithmically derived, their accuracies and precisions are highly varying with the camera networks, scene complexity, and dense image matching (DIM) algorithms, and there is no standard error metric to determine per-point errors. The theory of internal reliability of camera networks has been well studied through first-order error estimation of Bundle Adjustment (BA), which is used to understand the errors of 3D points assuming known measurement errors. However, the measurement errors of the DIM algorithms are intricate to an extent that every single point may have its error function determined by factors such as pixel intensity, texture entropy, and surface smoothness. Despite the complexity, there exist a few common metrics that may aid the process of estimating the posterior reliability of the derived points, especially in a multi-view stereo (MVS) setup when redundancies are present. In this paper, by using an aerial oblique photogrammetric block with LiDAR reference data, we analyze several internal matching metrics within a common MVS framework, including statistics in ray convergence, intersection angles, DIM energy, etc.