Our objective is language-based search of large-scale image and video datasets. For this
task, the approach that consists of independently mapping text and vision to a joint embedding …

… to the HCI alt-text dataset [8] for our evaluation. It consists of 3386 scientific figures with alt-text
descriptions extracted from publications on HumanComputer Interaction and accessibility. …

… 3) The recent work CLIP [43] provides a stronger vision encoder and we also evaluate the
performance based on it. It is shown that the CLIP’s weights greatly improve the performance …

… In essence, we solve a typical computer vision problem with … the images on the Web have
an HTML ALT caption). Today, it is the … Rather than designing a computer vision algorithm that …

… alternative text. To ameliorate this problem, we introduce WebInSight, a sysн tem that
automatically creates and inserts alternative text … image using computer vision techniques is …

… Enterprise Search, AI and Computer Vision to construct a single … Fourth, applying Computer
Vision techniques to extract … Figure 7, that utilizes Computer Vision to detect and extract …

… modal representations on image-text pairs are becoming popular for vision-language tasks.
… text features as input to the model to be pre-trained and use self-attention to learn image-text …

… convenient alternative method for their current meter reading updating system. The proposed
solution is to use computer vision techniques for capturing readings. One of the alternative …

… This framework overturned many computer science fields, like Computer Vision and Natural
Language Processing, obtaining astonishing results. Nevertheless, many challenges are …

… Inspired by the success of Transformers in NLP [20, 76], the Vision Transformer (ViT) [22]
was recently proposed and has been successfully applied to multiple computer vision tasks, …