CloudCV: Large-Scale Distributed Computer Vision as a Cloud Service

By Harsh Agrawal, Clint Solomon Mathialagan, Yash Goyal, Neelima Chavali, Prakriti Banik, Akrit Mohapatra, Ahmed A. A. Osman, Dhruv Batra. 2015

51 citations recorded by the source corpus.

We are witnessing a proliferation of massive visual data. Unfortunately scaling existing computer vision algorithms to large datasets leaves researchers repeatedly solving the same algorithmic, logistical, and infrastructural problems. Our goal is to democratize computer vision; one should not have to be a computer vision, big data and distributed computing expert to have access to state-of-the-art distributed computer vision algorithms. We present CloudCV, a comprehensive system to provide access to state-of-the-art distributed computer vision algorithms as a cloud service through a Web Interface and APIs.

Subfields: Computer Vision and Pattern Recognition, Electrical and Electronic Engineering, Radiology, Nuclear Medicine and Imaging, ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Electrical and Electronic Engineering

In Radiology, Nuclear Medicine and Imaging

In ML Systems & Hardware

Citations in the corpus

VQA: Visual Question Answering (2015, methodology): “nd language-alone baselines. As a general trend, results on multiple-choice are better than open-ended. All methods are signiﬁcantly worse than human performance. Our VQA demo is available on CloudCV [1] –http://cloudcv. org/vqa. This will be updated with newer models as we develop them. To gain fur”
Grad-CAM: Why did you say that? (2016, methodology): “We have released code, an online demo hosted on CloudCV [1], and the full paper [8].”
Towards Transparent AI Systems: Interpreting Visual Question Answering Models (2016, background): “We found that even without 1Demo available here: http://cloudcv.org/vqa/ (Agrawal et al., 2015) explicit attention mechanisms, VQA models may sometimes be implicitly attending to relevant regions in the image, and often to appropriate words in the question.”
Towards Open World Recognition (2014, methodology): “;;L;I] with: 1.A multi-class open set recognition function F(x) : Rd 7!N using a vector function ’(x) of iper-class measurable recognition functions f i(x), also using a novelty detector (’) : Ri 7![0;1]. We require the per class recognition functions f i(x) 2H: Rd 7!R for i2K t to be open set recog”
Understanding Locally Competitive Networks (2014, methodology): “For the experiments in this section, the penultimate-layer activations obtained using this model were downloaded from CloudCV [22].”
VIP: Finding important people in images (2015, background): “An online demo of our system available at [1, 3] does not use this feature.”
The Extreme Value Machine (2015, background): “DeCAF features [7] on ILSVRC2014 [25], courtesy of [1], which are derived from a truncated form of AlexNet [17].”
Semantic Aware Video Transcription Using Random Forest Classifiers (2014, methodology): “Part of the detector models were downloaded from [2].”