Deep Hough Transform for Semantic Line Detection

Zhao, Kai; Han, Qi; Zhang, Changbin; Xu, Jun; Cheng, Ming–Ming

doi:10.1109/tpami.2021.3077129

Cited by 92 publications

(51 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here, we used HT to extract features from the word images; more specifically, the vertical segments of the word images, obtained by the process as described in Section 3.2 . We should mention that many researchers have used HT to extract features for several image processing and pattern recognition tasks, such as finding strokes in geoscientific images [ 48 ], mammogram classification for early detection of breast cancer [ 49 ], face recognition [ 50 ], contextual line feature extraction for semantic line detection [ 51 ], detection of electric power bushings from infrared images [ 52 ], and many more. In general, in such applications, HT is used to find the straight lines in the image space.…”

Section: Present Workmentioning

confidence: 99%

Hough Transform-Based Angular Features for Learning-Free Handwritten Keyword Spotting

Kundu

Malakar²,

Geem

et al. 2021

Sensors

View full text Add to dashboard Cite

Handwritten keyword spotting (KWS) is of great interest to the document image research community. In this work, we propose a learning-free keyword spotting method following query by example (QBE) setting for handwritten documents. It consists of four key processes: pre-processing, vertical zone division, feature extraction, and feature matching. The pre-processing step deals with the noise found in the word images, and the skewness of the handwritings caused by the varied writing styles of the individuals. Next, the vertical zone division splits the word image into several zones. The number of vertical zones is guided by the number of letters in the query word image. To obtain this information (i.e., number of letters in a query word image) during experimentation, we use the text encoding of the query word image. The user provides the information to the system. The feature extraction process involves the use of the Hough transform. The last step is feature matching, which first compares the features extracted from the word images and then generates a similarity score. The performance of this algorithm has been tested on three publicly available datasets: IAM, QUWI, and ICDAR KWS 2015. It is noticed that the proposed method outperforms state-of-the-art learning-free KWS methods considered here for comparison while evaluated on the present datasets. We also evaluate the performance of the present KWS model using state-of-the-art deep features and it is found that the features used in the present work perform better than the deep features extracted using InceptionV3, VGG19, and DenseNet121 models.

show abstract

Section: Present Workmentioning

confidence: 99%

Hough Transform-Based Angular Features for Learning-Free Handwritten Keyword Spotting

Kundu

Malakar²,

Geem

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…For the application of semantic lines or horizon detection in natural scenes, Lee et al [13] proposed the VGG16-based semantic line network (SLNet) with line pooling layers, which combines line detection as a multitask loss of classification and regression. The deep Hough transform method by Zhao et al [14] incorporates the Hough transform into a one-shot end-to-end learning pipeline by using a CNN encoder with feature pyramids for feature extraction and performing the line detection in Hough space. Nguyen et al [15] transferred the ideas from object detection to design the LS-Net for power line detection that uses a CNN with two heads: one for classification and the other for line regression.…”

Section: Segmentation and Detection Of Linesmentioning

confidence: 99%

ChainLineNet: Deep-Learning-Based Segmentation and Parameterization of Chain Lines in Historical Prints

et al. 2021

View full text Add to dashboard Cite

The paper structure of historical prints is sort of a unique fingerprint. Paper with the same origin shows similar chain line distances. As the manual measurement of chain line distances is time consuming, the automatic detection of chain lines is beneficial. We propose an end-to-end trainable deep learning method for segmentation and parameterization of chain lines in transmitted light images of German prints from the 16th Century. We trained a conditional generative adversarial network with a multitask loss for line segmentation and line parameterization. We formulated a fully differentiable pipeline for line coordinates’ estimation that consists of line segmentation, horizontal line alignment, and 2D Fourier filtering of line segments, line region proposals, and differentiable line fitting. We created a dataset of high-resolution transmitted light images of historical prints with manual line coordinate annotations. Our method shows superior qualitative and quantitative chain line detection results with high accuracy and reliability on our historical dataset in comparison to competing methods. Further, we demonstrated that our method achieves a low error of less than 0.7 mm in comparison to manually measured chain line distances.

show abstract

“…Hough transform [24] converts curves in Euclidean space into points in parameter space, where a voting mechanism is used to detect the features of a given curve. For the binary image T, the progressive probabilistic Hough transform (PPHT) [25] can be applied to eliminate the influence of interference and obtain the straight-line segment describing the beacon.…”

Section: Extraction Of Beacon Linear Featuresmentioning

confidence: 99%

State Estimation of Axisymmetric Target Based on Beacon Linear Features and View Relation

Cao

Peng

Liu

2021

Sensors

View full text Add to dashboard Cite

In order to realize state estimation for axisymmetric targets and improve the accuracy and robustness of state estimation, a state estimation method for axisymmetric targets based on beacon linear features and view relation is proposed in this paper. The depth camera is used to collect the image and depth information of the object, and the features of the beacon line are extracted by the thinning process and Hough transform. Then, the rotation matrix model based on view relation is constructed to solve the target state. Finally, an axisymmetric shore power plug is taken as the experimental object and the L–V (linear features and view relation) state estimation method is compared with the C–H and C–IPPE state estimation methods. The experimental results show that the L–V state estimation method has higher accuracy and robustness.

show abstract

Deep Hough Transform for Semantic Line Detection

Cited by 92 publications

References 69 publications

Hough Transform-Based Angular Features for Learning-Free Handwritten Keyword Spotting

Hough Transform-Based Angular Features for Learning-Free Handwritten Keyword Spotting

ChainLineNet: Deep-Learning-Based Segmentation and Parameterization of Chain Lines in Historical Prints

State Estimation of Axisymmetric Target Based on Beacon Linear Features and View Relation

Contact Info

Product

Resources

About