OPTIMOL: Automatic Online Picture Collection via Incremental Model Learning

By Li-Jia Li, Li Fei-Fei. 2007

494 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

Learning Visual Features from Large Weakly Supervised Data (2015, background): “In particular, Li and Fei-Fei [27] present a hierarchical topic model that performs simultaneous dataset construction and incremental learning of object recognition models.”
The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (2015, methodology): “4 is inspired by methods that propose to learn from web data [13, 8, 9, 60, 40, 29, 16] or reason about label noise [34, 57, 52, 47, 38].”
Towards Open World Recognition (2014, methodology): “However, the multi-class incremental learning methods and other incremental classifiers [4, 36, 43, 21] are incremental in terms of additional training samples but not additional training categories.”
Data Distillation: Towards Omni-Supervised Learning (2017, methodology): “Among semi-supervised methods, our method is most related to self-training, a strategy in which a model’s predictions on unlabeled data are used to train itself [36, 48, 43, 33, 22, 46, 5, 21].”
Max-margin additive classifiers for detection (2009, background): “This complements a range of dataset collection efforts from semi-automatic [2, 25, 34, 5, 31] with 10,000-600,000+ images to fully manual but unpaid [32] with 50,000+ labeled objects to more traditional datasets [9, 13].”
Well Begun Is Half Done: Generating High-Quality Seeds for Automatic Image Dataset Construction from Web (2014, methodology): “The values of OPTIMOL are from [14,13], and the seeds' amount K in ELDA is set to be the same with ours.”
Towards total scene understanding: Classification, annotation and segmentation in an automatic framework (2009, methodology): “Finally, our approach is inspired by earlier approaches of learning from noisy internet data such as [2, 19, 10, 11, 29].”
NEIL: Extracting Visual Knowledge from Web Data (2013, methodology): “An alternative approach is to use visual recognition for extracting these datasets automatically from the Internet [23] [34] [36].”
Learning Models for Object Recognition from Natural Language Descriptions (2009, methodology): “Related work. A number of authors have investigated using single keyword searches to find example training images from the world wide web [2, 10, 17 , 19], and refining results using vision.”
Generalized Out-of-Distribution Detection: A Survey (2021, unknown): “Formally, OSR deals with the case where OOD samples during testing have semantic shift, i.e., P ( Y ) (cid:3)= P (cid:2) ( Y ) .”
Unbiased Metric Learning: On the Utilization of Multiple Datasets and Web Images for Softening Bias (2013, background): “Therefore, if used as training set, in order to get comparable results, extra labor is needed to remove noise and build a robust learning algorithm [8][10].”
Learning realistic human actions from movies (2008, unknown): “Our script-based annotation of human actions is similar in spirit to several recent papers using textual information for automatic image collection from the web [10, 14] and automatic naming of characters in images [1] and videos [4].”
Automatic Image Dataset Construction from Click-through Logs Using Deep Neural Network (2015, background): “These labelled images can be manually labelled [21] or top-ranked images from image search engines [2, 7, 17].”
Learning Object Categories From Internet Image Searches (2010, methodology): “Li et al. [27] show how methods, similar to those introduced in this paper, can be used in incremental fashion to compile a data set of a desired class from the Internet.”
Generating Diverse Image Datasets with Limited Labeling (2016, background): “To address the issues stated above and inspired by streams of images available on the web, there has been lot of recent interest in learning directly from noisy web data [20, 28, 18, 4, 14, 15, 5] or automatically curating web images for creating a dataset [33, 23, 31, 6, 1].”