Publications by Lorenzo Baraldi

Image captioning, like many tasks involving vision and language, currently relies on Transformer-based architectures for extracting the semantics in an image and translating it into linguistically coherent descriptions. Although successful, the attention operator only considers a weighted summation of projections of the current input sample, therefore ignoring the relevant semantic information which can come from the joint observation of other samples. In this paper, we devise a network which can perform attention over activations obtained while processing other training samples, through a prototypical memory model. Our memory models the distribution of past keys and values through the definition of prototype vectors which are both discriminative and compact. Experimentally, we assess the performance of the proposed model on the COCO dataset, in comparison with carefully designed baselines and state-of-the-art approaches, and by investigating the role of each of the proposed components. We demonstrate that our proposal can increase the performance of an encoder-decoder Transformer by 3.7 CIDEr points both when training in cross-entropy only and when fine-tuning with self-critical sequence training. Source code and trained models are available at: https://github.com/aimagelab/PMA-Net.

Image-text matching is gaining a leading role among tasks involving the joint understanding of vision and language. In literature, this task is often used as a pre-training objective to forge architectures able to jointly deal with images and texts. Nonetheless, it has a direct downstream application: cross-modal retrieval, which consists in finding images related to a given query text or vice-versa. Solving this task is of critical importance in cross-modal search engines. Many recent methods proposed effective solutions to the image-text matching problem, mostly using recent large vision-language (VL) Transformer networks. However, these models are often computationally expensive, especially at inference time. This prevents their adoption in large-scale cross-modal retrieval scenarios, where results should be provided to the user almost instantaneously. In this paper, we propose to fill in the gap between effectiveness and efficiency by proposing an ALign And DIstill Network (ALADIN). ALADIN first produces high-effective scores by aligning at fine-grained level images and texts. Then, it learns a shared embedding space – where an efficient kNN search can be performed – by distilling the relevance scores obtained from the fine-grained alignments. We obtained remarkable results on MS-COCO, showing that our method can compete with state-of-the-art VL Transformers while being almost 90 times faster. The code for reproducing our results is available at https://github.com/mesnico/ALADIN.

Handwritten Text Recognition (HTR) in free-layout pages is a challenging image understanding task that can provide a relevant boost to the digitization of handwritten documents and reuse of their content. The task becomes even more challenging when dealing with historical documents due to the variability of the writing style and degradation of the page quality. State-of-the-art HTR approaches typically couple recurrent structures for sequence modeling with Convolutional Neural Networks for visual feature extraction. Since convolutional kernels are defined on fixed grids and focus on all input pixels independently while moving over the input image, this strategy disregards the fact that handwritten characters can vary in shape, scale, and orientation even within the same document and that the ink pixels are more relevant than the background ones. To cope with these specific HTR difficulties, we propose to adopt deformable convolutions, which can deform depending on the input at hand and better adapt to the geometric variations of the text. We design two deformable architectures and conduct extensive experiments on both modern and historical datasets. Experimental results confirm the suitability of deformable convolutions for the HTR task.

Describing images in natural language is a fundamental step towards the automatic modeling of connections between the visual and textual modalities. In this paper we present CaMEL, a novel Transformer-based architecture for image captioning. Our proposed approach leverages the interaction of two interconnected language models that learn from each other during the training phase. The interplay between the two language models follows a mean teacher learning paradigm with knowledge distillation. Experimentally, we assess the effectiveness of the proposed solution on the COCO dataset and in conjunction with different visual feature extractors. When comparing with existing proposals, we demonstrate that our model provides state-of-the-art caption quality with a significantly reduced number of parameters. According to the CIDEr metric, we obtain a new state of the art on COCO when training without using external data. The source code and trained models will be made publicly available at: https://github.com/aimagelab/camel.

Image-based virtual try-on has recently gained a lot of attention in both the scientific and fashion industry communities due to its challenging setting and practical real-world applications. While pure convolutional approaches have been explored to solve the task, Transformer-based architectures have not received significant attention yet. Following the intuition that self- and cross-attention operators can deal with long-range dependencies and hence improve the generation, in this paper we extend a Transformer-based virtual try-on model by adding a dual-branch collaborative module that can exploit cross-modal information at generation time. We perform experiments on the VITON dataset, which is the standard benchmark for the task, and on a recently collected virtual try-on dataset with multi-category clothing, Dress Code. Experimental results demonstrate the effectiveness of our solution over previous methods and show that Transformer-based architectures can be a viable alternative for virtual try-on.

Embodied agents, trained to explore and navigate indoor photorealistic environments, have achieved impressive results on standard datasets and benchmarks. So far, experiments and evaluations have involved domestic and working scenes like offices, flats, and houses. In this paper, we build and release a new 3D space with unique characteristics: the one of a complete art museum. We name this environment ArtGallery3D (AG3D). Compared with existing 3D scenes, the collected space is ampler, richer in visual features, and provides very sparse occupancy information. This feature is challenging for occupancy-based agents which are usually trained in crowded domestic environments with plenty of occupancy information. Additionally, we annotate the coordinates of the main points of interest inside the museum, such as paintings, statues, and other items. Thanks to this manual process, we deliver a new benchmark for PointGoal navigation inside this new space. Trajectories in this dataset are far more complex and lengthy than existing ground-truth paths for navigation in Gibson and Matterport3D. We carry on extensive experimental evaluation using our new space for evaluation and prove that existing methods hardly adapt to this scenario. As such, we believe that the availability of this 3D model will foster future research and help improve existing solutions.

Image Captioning is the task of translating an input image into a textual description. As such, it connects Vision and Language in a generative fashion, with applications that range from multi-modal search engines to help visually impaired people. Although recent years have witnessed an increase in accuracy in such models, this has also brought increasing complexity and challenges in interpretability and visualization. In this work, we focus on Transformer-based image captioning models and provide qualitative and quantitative tools to increase interpretability and assess the grounding and temporal alignment capabilities of such models. Firstly, we employ attribution methods to visualize what the model concentrates on in the input image, at each step of the generation. Further, we propose metrics to evaluate the temporal alignment between model predictions and attribution scores, which allows measuring the grounding capabilities of the model and spot hallucination flaws. Experiments are conducted on three different Transformer-based architectures, employing both traditional and Vision Transformer-based visual features.

Exploration of indoor environments has recently experienced a significant interest, also thanks to the introduction of deep neural agents built in a hierarchical fashion and trained with Deep Reinforcement Learning (DRL) on simulated environments. Current state-of-the-art methods employ a dense extrinsic reward that requires the complete a priori knowledge of the layout of the training environment to learn an effective exploration policy. However, such information is expensive to gather in terms of time and resources. In this work, we propose to train the model with a purely intrinsic reward signal to guide exploration, which is based on the impact of the robot’s actions on its internal representation of the environment. So far, impact-based rewards have been employed for simple tasks and in procedurally generated synthetic environments with countable states. Since the number of states observable by the agent in realistic indoor environments is non-countable, we include a neural-based density model and replace the traditional count-based regularization with an estimated pseudo-count of previously visited states. The proposed exploration approach outperforms DRL-based competitors relying on intrinsic rewards and surpasses the agents trained with a dense extrinsic reward computed with the environment layouts. We also show that a robot equipped with the proposed approach seamlessly adapts to point-goal navigation and real-world deployment.

Vision Transformers (ViT) and other Transformer-based architectures for image classification have achieved promising performances in the last two years. However, ViT-based models require large datasets, memory, and computational power to obtain state-of-the-art results compared to more traditional architectures. The generic ViT model, indeed, maintains a full-length patch sequence during inference, which is redundant and lacks hierarchical representation. With the goal of increasing the efficiency of Transformer-based models, we explore the application of a 2D max-pooling operator on the outputs of Transformer encoders. We conduct extensive experiments on the CIFAR-100 dataset and the large ImageNet dataset and consider both accuracy and efficiency metrics, with the final goal of reducing the token sequence length without affecting the classification performance. Experimental results show that bidimensional downsampling can outperform previous classification approaches while requiring relatively limited computation resources.