Dataset Information

Blind Users Accessing Their Training Images in Teachable Object Recognizers.

ABSTRACT: Teachable object recognizers provide a solution for a very practical need for blind people - instance level object recognition. They assume one can visually inspect the photos they provide for training, a critical and inaccessible step for those who are blind. In this work, we engineer data descriptors that address this challenge. They indicate in real time whether the object in the photo is cropped or too small, a hand is included, the photos is blurred, and how much photos vary from each other. Our descriptors are built into open source testbed iOS app, called MYCam. In a remote user study in (N = 12) blind participants' homes, we show how descriptors, even when error-prone, support experimentation and have a positive impact in the quality of training set that can translate to model performance though this gain is not uniform. Participants found the app simple to use indicating that they could effectively train it and that the descriptors were useful. However, many found the training being tedious, opening discussions around the need for balance between information, time, and cognitive load.

SUBMITTER: Hong J

PROVIDER: S-EPMC10008526 | biostudies-literature | 2022 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Blind Users Accessing Their Training Images in Teachable Object Recognizers.

Hong Jonggi J Gandhi Jaina J Mensah Ernest Essuah EE Zeraati Farnaz Zamiri FZ Jarjue Ebrima Haddy EH Lee Kyungjun K Kacorri Hernisa H

ASSETS. Annual ACM Conference on Assistive Technologies 20221022

Teachable object recognizers provide a solution for a very practical need for blind people - instance level object recognition. They assume one can visually inspect the photos they provide for training, a critical and inaccessible step for those who are blind. In this work, we engineer data descriptors that address this challenge. They indicate in real time whether the object in the photo is cropped or too small, a hand is included, the photos is blurred, and how much photos vary from each other ...[more]

PMID: 36916963

Similar Datasets

Project description:BackgroundOne in ten people aged between 55 to 74 years have a significant hearing impairment in their better hearing ear (as defined by audiometric hearing thresholds). However, it is becoming increasingly clear that the challenges faced by older listeners cannot be explained by the audiogram alone. The ability for people with hearing loss to use cognition to support speech perception allows for compensation of the degraded auditory input. This in turn offers promise for new cognitive-based rehabilitative interventions. Working memory is known to be highly associated with language comprehension and recent evidence has shown significant generalization of learning from trained working memory tasks to improvements in sentence-repetition skills of children with severe to profound hearing loss. This evidence offers support for further investigation into the potential benefits of working memory training to improve speech perception abilities in other hearing impaired populations.Methods/designThis is a double-blind randomized active controlled trial aiming to assess whether a program of working memory training results in improvements in untrained measures of cognition, speech perception and self-reported hearing abilities in adult hearing aid users aged (50 to 74 years) with mild-to-moderate hearing loss hearing aid users, compared with an active control group who receive a placebo version of the working memory training program.DiscussionThe present study aims to generate high-quality preliminary evidence for the efficacy of working memory training for adult hearing aid users with mild-to-moderate sensorineural hearing loss who are existing hearing aid users. This trial addresses a number of gaps in the published literature assessing training interventions for people with hearing loss, and in the general literature surrounding working memory training, such as the inclusion of an active control group, participant and tester blinding, and increased transparency in reporting.Trial registrationClinicalTrials.gov identifier: NCT01892007. Date of registration: 27 June 2013.

Dataset Information

Blind Users Accessing Their Training Images in Teachable Object Recognizers.

Publications

Blind Users Accessing Their Training Images in Teachable Object Recognizers.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets