Content-based image retrieval using feature-fusion of GroupNormalized-Inception-Darknet-53 features and handcraft features

Pathak, Debanjan; Raju, U. S. N.

doi:10.1016/j.ijleo.2021.167754

Cited by 21 publications

(8 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Gaussian distribution parameter σ in this function plays a decisive role in the width of the Gaussian filter. In the plane image, there is a large correlation between the RGB color space channels, which is easy to cause the image correction deviation based on the reflection coefficient component; so, the HSV color model (i.e., hue (H), saturation (S ), and lightness (V)) is used to collect illumination information of images [24,25]. It is corrected for the V channel (i.e., lightness) in HSV space by gamma correction [26].…”

Section: Optimization Model For Color Enhancement Processing Of Plane...mentioning

confidence: 99%

Optimization of Color Enhancement Processing for Plane Images Based on Computer Vision

Jiao

2022

Journal of Sensors

View full text Add to dashboard Cite

As the main carrier of information transmission, plane images play an important role in the current network era. In order to enhance the color of the plane image, optimize the image effect, and solve the problem of image distortion with large color difference, a color enhancement processing optimization method is proposed based on the technical support of computer vision. According to computer vision theory, a computer vision model for adjusting perceived color and brightness is constructed. Combining it with the bilateral filtering algorithm, a color enhancement processing optimization model is obtained, which consists of three main stages: illumination information parameter estimation and image color correction, reflection coefficient parameter estimation and image color correction, and adaptive filtering. By estimating illumination information parameters and reflection coefficient parameters, the secondary gamma correction of image color is realized. After being processed by the bilateral filtering algorithm, the final optimized image is obtained. Verified by the subjective and objective evaluation results, the positive correlation index values of this method are all high, 8.0246, 16.4526, 0.9037, and 15.0246, respectively, and the negative correlation index value is low, 49.4169. It is proved that the proposed method can effectively suppress the halo and artifacts and obtain the image color quality that satisfies the visual perception. It not only enhances the image color but also retains the details to the greatest extent.

show abstract

Section: Optimization Model For Color Enhancement Processing Of Plane...mentioning

confidence: 99%

Optimization of Color Enhancement Processing for Plane Images Based on Computer Vision

Jiao

2022

Journal of Sensors

View full text Add to dashboard Cite

show abstract

“…As RGB handcrafted features an improved version of DDBTC is used. 68 In the traditional DDBTC bitmap image is calculated from the gray version of RGB input image. This gray image was constructed using Equation (3), which is the equal ratio of R, G, and B component.…”

Section: Methodsmentioning

confidence: 99%

“…Feature vector extracted from both HSI and RGB color space is used as handcraft feature. As RGB handcrafted features an improved version of DDBTC is used 68 . In the traditional DDBTC bitmap image is calculated from the gray version of RGB input image.…”

Section: Methodsmentioning

confidence: 99%

Content‐based image retrieval for super‐resolutioned images using feature fusion: Deep learning and hand crafted

Pathak

Raju

2022

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

An innovative image retrieval agenda by concatenating deep learning features from GoogleNet and low‐level features from HSI and RGB color space is proposed in this article. Most of the CNN features suffer from loss of information due to image resize as a pre‐processing stage. To reduce this information loss super‐resolution technic is used for resizing images. An improved form of dot‐diffused block truncation coding is used for extracting RGB handcraft features. To discover the interdependencies between color and intensity component of an image, interchannel voting between hue, saturation, and intensity component is calculated as a color feature in HSI space. Histogram of orientated gradient (HOG) feature is used as shape feature. Five standard performance parameters, average precision rate, average recall rate, F‐Measure, Average Normalized Modified Retrieval Rank, and Total Minimum Retrieval Epoch, are applied on nine image datasets: Corel‐1K, Corel‐5K, Corel‐10K, VisTex, STex, ColorBrodatz and three subsets of ImageNet dataset for evaluation process of proposed method. For all dataset the best performance is achieved by the proposed method with respect to all performance parameters.

show abstract

“…Aiming at the problems of low image SNR caused by the influence of illumination in the actual scene of the substation and weak spatial generalization caused by image strabismus and different distribution of knob features, we innovatively combine the YOLOv4 [15][16][17] target detection algorithm with the Darknet53 [36][37][38] feature extraction network, the DUC module, and the DSNT module to form a knob gear detection network. We conduct experiments on the knob image data collected in the substation, comparing it with other researchers' knob gear detection methods.…”

Section: Related Workmentioning

confidence: 99%

Recognition Method of Knob Gear in Substation Based on YOLOv4 and Darknet53-DUC-DSNT

Qin

Hua

Sun

et al. 2022

Sensors

View full text Add to dashboard Cite

When unattended substations are popular, the knob is a vital monitoring object for unattended substations. However, in the actual scene of the substation, the recognition method of a knob gear has low accuracy. The main reasons are as follows. Firstly, the SNR of knob images is low due to the influence of lighting conditions, which are challenging to extract image features. Secondly, the image deviates from the front view affected by the shooting angle; that knob has a certain deformation, which causes the feature judgment to be disturbed. Finally, the feature distribution of each kind of knob is inconsistent, which interferes with image extraction features and leads to weak spatial generalization ability. For the above problems, we propose a three-stage knob gear recognition method based on YOLOv4 and Darknet53-DUC-DSNT models for the first time and apply key point detection of deep learning to knob gear recognition for the first time. Firstly, YOLOv4 is used as the knob area detector to find knobs from a picture of a cabinet panel. Then, Darknet53, which can extract features, is used as the backbone network for keypoint detection of knobs, combined with DUC structure to recover detailed information and DSNT structure to enhance feature extraction and improve spatial generalization ability. Finally, we obtained the knob gear by calculating the angle between the line of the rotating center point and the pointing point and horizontal direction. The experimental results show that this method effectively solves the above problems and improves the performance of knob gear detection.

show abstract

Content-based image retrieval using feature-fusion of GroupNormalized-Inception-Darknet-53 features and handcraft features

Cited by 21 publications

References 49 publications

Optimization of Color Enhancement Processing for Plane Images Based on Computer Vision

Optimization of Color Enhancement Processing for Plane Images Based on Computer Vision

Content‐based image retrieval for super‐resolutioned images using feature fusion: Deep learning and hand crafted

Recognition Method of Knob Gear in Substation Based on YOLOv4 and Darknet53-DUC-DSNT

Contact Info

Product

Resources

About