An Approach to identify Captioning Keywords in an Image using LIME

Sahay, Siddharth; Omare, Nikita; Shukla, K. K.

doi:10.1109/icccis51004.2021.9397159

Cited by 14 publications

(2 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the use of a Gaussian (RBF) kernel, LIME assigns weights to each generated point. The size of the meaningful weights' circle around the red dot is determined by the kernel width kw option [47].…”

Section: Post Agnostic Models: Shap I Limementioning

confidence: 99%

Delving into Human Factors through LSTM by Navigating Environmental Complexity Factors within Use Case Points for Digital Enterprises

Rankovic,

Rankovic

2024

JTAER

View full text Add to dashboard Cite

Meeting customer requirements in software project management, even for large digital enterprises, proves challenging due to unpredictable human factors. It involves meticulous planning and environmental factor analysis, ultimately benefiting both companies and customers. This paper came as a natural extension of our previous work where we left ourselves curious about what impact environmental complexity factors (ECFs) have in a use case point (UCP) approach. Additionally, we wanted to possibly decrease the mean magnitude relative error (MMRE) with deep learning models such as long-short-term-memory (LSTM) and gradient recurrent unit (GRU). The data augmentation technique was used to artificially increase the number of projects, since in the industry world, digital enterprises are not keen to share their data. The LSTM model outperformed the GRU and XGBoost models, while the average MMRE in all phases of the experiment for all models achieved 4.8%. Moreover, the post-agnostic models showed the overall and individual impact of eight ECFs, where the third ECF “team experience” on a new project has been shown as the most influential one. Finally, it is important to emphasize that effectively managing human factors within ECFs in UCPs can have a significant impact on the successful completion of a project.

show abstract

Section: Post Agnostic Models: Shap I Limementioning

confidence: 99%

Delving into Human Factors through LSTM by Navigating Environmental Complexity Factors within Use Case Points for Digital Enterprises

Rankovic,

Rankovic

2024

JTAER

View full text Add to dashboard Cite

show abstract

“…In addition, many research has been conducted in this regard and concept was adopted to clarify why captioning system produced a specific description for a given image. For instance, [25] employed XAI to explain predictions of a captioning model by depicting a part of image corresponding to a particular word and showing why the model generated this word. Similarly, [26] added an explanatory layer to the stateof-the-art Show, "Attend and Tell" model by augmenting the attention mechanism using additional bottom-up features.…”

Section: B Explainable Image Captioningmentioning

confidence: 99%

Explainability for Medical Image Captioning

Beddiar

Oussalah

Seppänen

2022

2022 Eleventh International Conference on Image Processing Theory, Tools and Applications (IPTA)

View full text Add to dashboard Cite

Medical image captioning is the process of generating clinically significant descriptions to medical images, which has many applications among which medical report generation is the most frequent one. In general, automatic captioning of medical images is of great interest for medical experts since it offers assistance in diagnosis, disease treatment and automating the workflow of the health practitioners. Recently, many efforts have been put forward to obtain accurate descriptions but medical image captioning still provides weak and incorrect descriptions. To alleviate this issue, it is important to explain why the model produced a particular caption based on some specific features. This is performed through Artificial Intelligence Explainability (XAI), which aims to unfold the 'black-box' feature of deeplearning based models. We present in this paper an explainable module for medical image captioning that provides a sound interpretation of our attention-based encoder-decoder model by explaining the correspondence between visual features and semantic features. We exploit for that, self-attention to compute word importance of semantic features and visual attention to compute relevant regions of the image that correspond to each generated word of the caption in addition to visualization of visual features extracted at each layer of the Convolutional Neural Network (CNN) encoder. We finally evaluate our model using the ImageCLEF medical captioning dataset.

show abstract