Deep Learning Architectures for Computer Vision Applications: A Study

Bagi, Randheer; Dutta, Tanima; Gupta, Hari Prabhat

doi:10.1007/978-981-15-0694-9_56

Cited by 10 publications

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While strides have been taken toward achieving comprehensive image comprehension in CV, with impressive performance in tasks such as image classification, object detection, and image generation (Bagi et al, 2020), this progression aligns with the focus on low‐level features for tasks focused on perceptually bounded ground truths. Specifically, CV has had a strong focus on CNNs (LeCun et al, 2015) and other methods that target images' local‐level features, which have largely banked on the decontextualized nature of visual signals.…”

Section: The Technical Challenge Of Acmentioning

confidence: 98%

The wicked problem of naming the intangible: Abstract concepts, binary thinking, and computer vision labels

Martinez Pandiani

2024

Future Humanities

View full text Add to dashboard Cite

Delving into the intricate complexities of naming and categorizing the visual evocation of abstract concepts, this paper brings to light the limitations of relying on binary thinking to tackle these inherently intricate “wicked problems.” As computer vision applications rapidly expand, the pressing challenge of accurately labeling these abstract concepts in visual media comes into focus, necessitating a close examination of the interplay between visual data, nuanced cultural meanings, and artificial intelligence (AI). This work discusses the role these concepts play in automatic visual indexing, as well as the ways in which they expose how binary frameworks curtail technical performance and perpetuate power dynamics. To address this, the paper draws upon insights from recent cognitive neuroscience research and advocates for a more comprehensive, queer, and situated understanding of these concepts. This approach highlights the significance of humanistic and ethical perspectives in shaping the trajectory of AI development.

show abstract

Section: The Technical Challenge Of Acmentioning

confidence: 98%