AppTek’s Workbench Delivers 85% More Efficiency in Computer Vision and Speech Recognition Data Labeling Tasks
AppTek, a leader in Artificial Intelligence (AI), Machine Learning (ML), Automatic Speech Recognition (ASR), Neural Machine Translation (NMT), Text-to-Speech (TTS) and Natural Language Processing / Understanding (NLP/U) technologies, announced significant productivity gains through improvements to its Workbench Data Annotation and Labeling Platform. This is the result of ML-enabled automation for the production of multi-format, multi-domain, speech, audio, text, image, natural language and computer vision data, which are used to fuel mission critical AI models for federal and enterprise customers.
AppTek’s Workbench, used by its distributed workforce spanning across 70+ countries, serves as a secure, cloud-based data annotation and labeling platform designed to radically streamline the production of high-quality data sets. Recently, AppTek announced the expansion of Workbench to include rapid video labeling capabilities for computer vision (CV) models through automation afforded by cutting-edge detection and segmentation algorithms combined with NLU technologies for enhanced object detection. Workbench combines audio, text, image and video labeling processes, and features semi-automated speech and CV labeling to more efficiently produce a wide array of outputs. The resulting data is verified through a robust quality assurance process which involves automated confidence scores and human-in-the-loop validation for enterprise-ready results.
In head-to-head evaluations for video annotation tasks that used AppTek’s new semi-automated labeling capabilities, evaluators reported an 85% faster time-to-completion versus manual annotation tasks. Evaluators also recorded a reduction of 50% in the number of video frames that need to be analyzed per project and a 7x improvement in time to annotate objects per frame for instances where human-in-the loop editing was required. Audio evaluators also noted an impressive 75% productivity gain to fully segment, transcribe, annotate and label data for the training of speech recognition models. This is achieved through platform improvements, such as automatic time synchronization, automatic alignment, keyboard shortcuts, built-in tags and enforced validation rules to enhance the quality assurance process.
With these new audio and video labeling engines in place, AppTek can now offer its customers a suite of production-quality tested-and-validated AI data packages at a fraction of the cost and at rapid speed to market. The addition of AppTek’s new CV classifiers further expands the company’s data services portfolio to support a wide range of multi-format multi-domain outputs which include object detection, activity detection, scene detection, facial detection, video segmentation, rich metadata tagging, facial blurring, redaction of sensitive information, optical character recognition (OCR), and more. Workbench works across large video repositories and supports multi-modal annotation to allow customers to further explore the complex relationships between audio, video and text.
“We are thrilled with the new efficiencies and productivity gains offered by the semi-automation of audio and video labeling tasks inside the Workbench, ” said Katie Nguyen, SVP Data Operations. “Our portfolio of data sets and services continues to rapidly expand with new and diverse offerings for our customer base.”
“AppTek continues to deliver the most trusted, quality assured and scientifically validated data sets for the production of high-performing AI models” said AppTek CEO Mudar Yaghi. “And now we can deliver enterprise-grade multi-domain, multi-format data at just a fraction of the time and cost of manual-only offerings.”
For more information, visit http://www.apptek.com.
About AppTek
AppTek is a global leader in artificial intelligence (AI) and machine learning (ML) technologies for automatic speech recognition (ASR), neural machine translation (NMT), natural language processing/understanding (NLP/U) and text-to-speech (TTS) technologies. The AppTek platform delivers industry-leading, real-time streaming and batch technology solutions in the cloud or on-premises for organizations across a breadth of global markets such as media and entertainment, call centers, government, enterprise business, and more. Built by scientists and research engineers who are recognized among the best in the world, AppTek’s solutions cover a wide array of languages, dialects and channels. For more information, please visit http://www.apptek.com.
