Connecting Vision and Language plays an essential role in Generative Intelligence. For this reason, large research efforts have been devoted to image captioning, i.e. describing images with syntactically and semantically meaningful sentences. Starting from 2015 the task has generally been addressed with pipelines composed of a visual encoder and a language model for text generation. During these years, both components have evolved considerably through the exploitation of object regions, attributes, the introduction of multi-modal connections, fully-attentive approaches, and BERT-like early-fusion strategies. However, regardless of the impressive results, research in image captioning has not reached a conclusive answer yet. This work aims at providing a comprehensive overview of image captioning approaches, from visual encoding and text generation to training strategies, datasets, and evaluation metrics. In this respect, we quantitatively compare many relevant state-of-the-art approaches to identify the most impactful technical innovations in architectures and training strategies. Moreover, many variants of the problem and its open challenges are discussed. The final goal of this work is to serve as a tool for understanding the existing literature and highlighting the future directions for a research area where Computer Vision and Natural Language Processing can find an optimal synergy.

Deep-learning models for 3D point cloud semantic segmentation exhibit limited generalization capabilities when trained and tested on data captured with different sensors or in varying environments due to domain shift. Domain adaptation methods can be employed to mitigate this domain shift, for instance, by simulating sensor noise, developing domain-agnostic generators, or training point cloud completion networks. Often, these methods are tailored for range view maps or necessitate multi-modal input. In contrast, domain adaptation in the image domain can be executed through sample mixing, which emphasizes input data manipulation rather than employing distinct adaptation modules. In this study, we introduce compositional semantic mixing for point cloud domain adaptation, representing the first unsupervised domain adaptation technique for point cloud segmentation based on semantic and geometric sample mixing. We present a two-branch symmetric network architecture capable of concurrently processing point clouds from a source domain (e.g. synthetic) and point clouds from a target domain (e.g. real-world). Each branch operates within one domain by integrating selected data fragments from the other domain and utilizing semantic information derived from source labels and target (pseudo) labels. Additionally, our method can leverage a limited number of human point-level annotations (semi-supervised) to further enhance performance. We assess our approach in both synthetic-to-real and real-to-real scenarios using LiDAR datasets and demonstrate that it significantly outperforms state-of-the-art methods in both unsupervised and semi-supervised settings.

The semantic segmentation of underwater imagery is an important step in the ecological analysis of coral habitats. To date, scientists produce fine-scale area annotations manually, an exceptionally time-consuming task that could be efficiently automatized by modern CNNs. This paper extends our previous work presented at the 3DUW’19 conference, outlining the workflow for the automated annotation of imagery from the first step of dataset preparation, to the last step of prediction reassembly. In particular, we propose an ecologically inspired strategy for an efficient dataset partition, an over-sampling methodology targeted on ortho-imagery, and a score fusion strategy. We also investigate the use of different loss functions in the optimization of a Deeplab V3+ model, to mitigate the class-imbalance problem and improve prediction accuracy on coral instance boundaries. The experimental results demonstrate the effectiveness of the ecologically inspired split in improving model performance, and quantify the advantages and limitations of the proposed over-sampling strategy. The extensive comparison of the loss functions gives numerous insights on the segmentation task; the Focal Tversky, typically used in the context of medical imaging (but not in remote sensing), results in the most convenient choice. By improving the accuracy of automated ortho image processing, the results presented here promise to meet the fundamental challenge of increasing the spatial and temporal scale of coral reef research, allowing researchers greater predictive ability to better manage coral reef resilience in the context of a changing environment.