Humans upload over 1.8 billion digital images to the internet each day, yet the relationship between the images that a person shares with others and his/her psychological characteristics remains poorly understood. In the current research, we analyze the relationship between images, captions, and the latent demographic/psychological dimensions of personality and gender. We consider a wide range ...

Image captioning, a popular topic in computer vision, has achieved substantial progress in recent years. However, the distinctiveness of natural descriptions is often overlooked in previous work. It is closely related to the quality of captions, as distinctive captions are more likely to describe images with their unique aspects. In this work, we propose a new learning method, Contrastive Learn...

This project compares the accuracy of two automatic speech recognition (ASR) systems–Bing Speech and YouTube’s automatic captions–across gender, race and four dialects of American English. The dialects included were chosen for their acoustic dissimilarity. Bing Speech had differences in word error rate (WER) between dialects and ethnicities, but they were not statistically reliable. YouTube’s a...

The captions for Figure 2C,D have several minor errors numbering of the residues of Aβ. The corrected captions should read: (C) NMR-based structural model of Aβ1−40 fibrils showing central residues K16−D22 and A30−V36. (D) NMR-based structural model of Aβ1−42 fibrils showing central residues L17−I41. These errors are peripheral to this paper and do not affect the results or conclusions in any way.

We exploit the retrieval of visual information from biomedical scientific publication databases. Therefore, we consider the use of domain specific genres to automatically subdivide large image databases into smaller, consistent parts. Combination with Latent Semantic Indexing on the picture captions allows for efficient retrieval of images in specific categories. We demonstrate our approach on ...

This report presents our submission to the MS COCO Captioning Challenge 2015. The method uses Convolutional Neural Network activations as an embedding to find semantically similar images. From these images, the most typical caption is selected based on unigram frequencies. Although the method received low scores with automated evaluation metrics and in human assessed average correctness, it is ...

There were errors that occurred in the captions of Figures 2 and 3. The corrections of errors in these captions should be as follows. Figure 2: Duration of transient worsening during steroids pulse therapy. Most of the worsening occurs between 3 and 5 days in the treatment group and between 3 and 10 days in the control group. D: day. Figure 3: The effect results were shown in both groups with d...

Overview Our final project involves a do-it-yourself newscast. It requires a main feed that has a person standing in front of a distinguishable screen and preloaded pictures and corresponding captions. The user is also allowed basic control of the system including changing between different picture and captions and also making the picture become fullscreen to include more detail. This system al...

The automatic generation of image captions has received considerable attention. The problem of evaluating caption generation systems, though, has not been that much explored. We propose a novel evaluation approach based on comparing the underlying visual semantics of the candidate and ground-truth captions. With this goal in mind we have defined a semantic representation for visually descriptiv...

The rapid growth in the volume of visual information can make the task of finding and accessing visual information of interest, overwhelming for users. Semantic analysis of image captions can be used in conjunction with image retrieval systems (IMR) to retrieve selected images more precisely. To do this, we first exploit a Natural Language Processing (NLP) framework in order to extract concepts...