Learning to Count with CNN Boosting

By Elad Walach, Lior Wolf. European Conference on Computer Vision, 2016

316 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/806d7b97c3535a3c62ce243fe7008149062d14c1

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

Generating High-Quality Crowd Density Maps Using Contextual Pyramid CNNs (2017, methodology): “[36] proposed a layered approach of learning CNNs for crowd counting by iteratively adding CNNs where every new CNN is trained on residual error of the previous layer.”
A Survey of Recent Advances in CNN-based Single Image Crowd Counting and Density Estimation (2017, methodology): “…and counting, they however cover only traditional methods that use hand-crafted features and do not take into account the recent advancements driven primarily by CNN-based approaches [87, 39, 113, 11, 85, 97, 4, 98, 111, 107, 70, 88] and creation of new challenging crowd datasets [106, 107, 111].”
A comprehensive review on ensemble deep learning: Opportunities and challenges (2023, background): “In Walach and Wolf (2016) proposed boosting CNNs for object counting in images using different image datasets, namely mall crowd counting (Chen et al., 2012).”
Crowd Counting and Density Estimation by Trellis Encoder-Decoder Networks (2019, methodology): “Notably, most of these methods follow a patch-based counting mechanism [52, 44, 27, 1, 37, 21, 11, 2], where the full density map is obtained by concatenating discrete density patches.”
Bayesian Loss for Crowd Count Estimation With Point Supervision (2019, background): “…potentials to handle similar tasks in other domains, includ-∗ ing estimating the number of vehicles in trafﬁc congestion [29, 30, 56, 14, 28], counting the cells and bacteria from microscopic images [20, 42, 44, 45, 8], and animal crowd estimations for ecological survey [27, 1, 18], to name a few.”
A review of ensemble learning and data augmentation models for class imbalanced problems: combination, implementation and evaluation (2023, methodology): “It can also be used with deep learning models such as Convolutional Neural Networks [120] and other machine learning models[121].”
DecideNet: Counting Varying Density Crowds Through Attention Guided Detection and Density Estimation (2017, methodology): “This is achieved without using the ensemble scheme employed by the “MoCNN” and “Boosting CNN” methods.”
CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes (2018, methodology): “In the work presented by Walach and Wolf [36], a method is demonstrated with selective sampling and layered boosting.”
Leveraging Unlabeled Data for Crowd Counting by Learning to Rank (2018, methodology): “However the MSE of our method on UCF CC 50 dataset is worse then the state-of-the-art methods [33, 31], which means our methods works better in general but has more extreme outliers.”
CNN-Based cascaded multi-task learning of high-level prior and density estimation for crowd counting (2017, background): “[27] used CNNs with layered training approach.”
JHU-CROWD++: Large-Scale Crowd Counting Dataset and A Benchmark Method (2020, methodology): “Typically, most existing works ([4, 5, 6, 7, 22, 23, 24, 25, 26, 26, 30, 31]) have designed their networks by laying a strong emphasis on addressing large variations of scale in crowd images.”
Where are the Blobs: Counting by Localization with Point Supervision (2018, unknown): “ the full test set, whereas the other methods take the mean of ten random samples of the test set evaluation. Methods UCSD Mall ShanghaiTech B FCN-rLSTM [43] 1.54 - - MoCNN [45] - 2.75 - CNN-boosting [46] 1.10 2.01 - M-CNN [17] 1.07 - 26.4 CP-CNN [47] - - 20.1 CSRNet [44] 1.16 - 10.6 LC-FCN8 1.51 2.”
ADCrowdNet: An Attention-Injective Deformable Convolutional Network for Crowd Understanding (2018, methodology): “Crowd understanding by CNN: Inspired by the great success in visual classification and recognition, literature also focuses on the CNN-based approaches to predict crowd density map and count the number of crowds [32, 20, 16, 13, 23].”
Scale Aggregation Network for Accurate and Efficient Crowd Counting (2018, methodology): “In [29], Walach and Wolf made use of layered boosting and selective sampling methods to reduce the count estimation error.”