Online domain adaptation of a pre-trained cascade of classifiers

By Vidit Jain, Eric Learned-Miller. 2011

212 citations recorded by the source corpus.

How researchers see this paper

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

In Artificial Intelligence

Citations in the corpus

Unsupervised Model Adaptation for Continual Semantic Segmentation (2020, methodology): “This model adaptation setting been explored for non-deep models (Dredze and Crammer 2008; Jain and Learned-Miller 2011; Wu 2016), but these works cannot be extended to semantic segmentation tasks.”
Lifelong Generative Modeling (2017, background): “There are some exceptions to this where online learning is applied to the problem of domain adaptation, eg: [43, 40].”
Undoing the Damage of Dataset Bias (2012, methodology): “While most domain adaptation models need to be re-trained for every new domain, Jain and Learned-Miller [10] proposed an online method based on Gaussian process regression that rapidly adapts to the new domain without re-training.”
From Facial Parts Responses to Face Detection: A Deep Learning Approach (2015, methodology): “We compare Faceness-Net against all published methods [37, 23, 3, 35, 18, 20, 17, 28, 40, 13] in the FDDB.”
Learning to (Learn at Test Time): RNNs with Expressive Hidden States (2024, unknown): “In computer vision, the idea of learning at test time has been applied for decades to applications such as face detection [41], object detection [56], image super-resolution [68], and 3D reconstruction [53].”
Test-Time Training with Self-Supervision for Generalization under Distribution Shifts (2019, background): “Note that some of the previously covered papers (Hoffman et al., 2014; Jain & Learned-Miller, 2011; Mullapudi et al., 2018) use the term “online learning” outside of the learning theory setting, so the term can be overloaded.”
Perceptual Annotation: Measuring Human Vision to Improve Computer Vision (2014, result): “The perceptually annotated classifiers trained using the biologically-inspired features of Cox and Pinto [41] produce the best results compared to all prior published approaches reporting on this data set [37], [42], [43], [44].”
Joint Cascade Face Detection and Alignment (2014, result): “On FDDB dataset, we compare with all the published results in the platform [26,32,24,14,23,12,16,8,25,18,30,13,15].”
A Comprehensive Survey on Test-Time Adaptation Under Distribution Shifts (2023, unknown): “GPR [296] is one of the early works that only adjusts the network predictions instead of the pre-trained model.”
Local Transform Features and Hybridization for Accurate Face and Human Detection (2013, background): “See http://vis-www.cs.umass.edu/fddb/results.html. lower than those using the MIT+CMU database because the face images in the FDDB database have higher variations in pose, illumination, expression, and occlusion than those in the MIT+CMU database, 2) the detection rate of the proposed HYBRID method ”
Consistent video depth estimation (2020, background): “…different problem contexts: online update in visual tracking [Kalal et al. 2011; Ross et al. 2008], adapting object detectors from images to videos [Jain and Learned-Miller 2011; Tang et al. 2012], and learning video-specific features for person re-identification [Cinbis et al. 2011; Zhang et al.…”
Incremental Evolving Domain Adaptation (2016, methodology): “The online domain adaptation method [15] has a similar setting to [14], while the employed Gaussian Process Regression requires a large number of labeled samples from the target domain.”
Hand-Object Contact Consistency Reasoning for Human Grasps Generation (2021, methodology): “We then leverage the contact consistency between outputs of the two networks for adjusting the initial grasp, where the target contact map serves as a self-supervision signal. research has started tackling the problem by leveraging self-supervision at test-time [22, 42, 1, 44, 23, 34].”
Tent: Fully Test-Time Adaptation by Entropy Minimization (2021, background): “Test-time adaptation of predictions, through which harder and uncertain cases are adjusted based on easier and certain cases (Jain & Learned-Miller, 2011), provides inspiration for certainty-based model adaptation schemes like our own.”
A survey on face detection in the wild: Past, present and future (2015, background): “Some early work has been conducted in this area [219, 220, 202, 203] and we strongly believe that this is a great direction for future work.”
DLID: Deep Learning for Domain Adaptation by Interpolating between Domains (2013, background): “Such issues of a mismatched train and test sets occur not only in the field of Computer Vision (Duan et al., 2009; Jain & Learned-Miller, 2011; Wang & Wang, 2011), but also in Natural Language Processing (Blitzer et al., 2006; 2007; Glorot et al., 2011), and Automatic Speech Recognition (Leggetter…”