Optimized Block-Based Connected Components Labeling With Decision Trees

Grana, Costantino; Borghesani, Daniele; Cucchiara, Rita

doi:10.1109/tip.2010.2044963

Cited by 155 publications

(181 citation statements)

References 25 publications

Supporting

Mentioning

176

Contrasting

Unclassified

Order By: Relevance

“…The next step is the creation of area kernels, i.e., connected regions of common bathymorphon class. The area kernels are created by applying the connected components labeling algorithm (specifically, the Block Based with Decision Trees algorithm described in [36]) to the seafloor form-classified grid creating a bathy-morphometric map. This step of the algorithm scans the created map with bathymorphon classes, and groups its nodes into components based on node connectivity: all nodes in a connected component share same bathymorphon class and are connected with each other.…”

Section: Area Kernels Based On Landform Classificationmentioning

confidence: 99%

A Bathymetry- and Reflectivity-Based Approach for Seafloor Segmentation

2018

View full text Add to dashboard Cite

A robust and flexible technique to segment seafloor acoustic mapping data by analyzing co-located bathymetric digital elevation models and acoustic backscatter mosaics is presented. The algorithm first uses principles of topographic openness, pattern recognition, and texture classification to identify geomorphic elements of the seafloor or "area kernels", and then derives the final seafloor segmentation by merging or splitting the kernels based on principles of similarity and multi-modality. The output is a collection of homogeneous, non-overlapping seafloor segments of consistent morphology and acoustic backscatter texture. Each labeled segment is enriched by a list of derived, physically-meaningful attributes that can be used for subsequent task-specific analysis.

show abstract

Section: Area Kernels Based On Landform Classificationmentioning

confidence: 99%

A Bathymetry- and Reflectivity-Based Approach for Seafloor Segmentation

2018

View full text Add to dashboard Cite

show abstract

“…Block-based labeling and efficient block processing with decision table were proposed in [2,3]. Block equivalence method based on "scan mask" was proposed [10].…”

Section: Related Workmentioning

confidence: 99%

Gpu-Based Object Identification in Large-Scale Images for Real-Time Radar Signal Analysis

Shioya¹,

Yamamoto²

2016

IJNCAA

View full text Add to dashboard Cite

As the computing power of processors is being drastically improved, the sizes of image data for various applications are also increasing. However, CCL cannot be easily implemented in a parallel fashion because the connected pixels can be found basically only by graph traversal. In this paper, we propose a GPU-based efficient algorithm for object identification in large-scale images and the performance of the proposed method is compared with that of the most commonly used method implemented with OpenCV libraries. The experimental results show that the proposed method outperforms the reference method when the pixel density is below 0.7. Object identification in image data is the fundamental operation and rapid computation is highly requested as the sizes of the currently available image data rapidly increase. The method proposed in this paper is applied to the analysis of radar signals. Since the radar signals contains complex and large data, the efficiency of the method is crucial when realtime analysis is required. The experimental results show the proposed method can be a good solution to the object identification in largescale image data such as radar signals.

show abstract

“…With a focus on natural requirements such as consistency and conservativeness, incremental learning is analyzed both for learning from positive examples and for learning from positive and negative examples. In [19] authors introduced a novel form of decision tables, namely OR-Decision Tables, which allow including the representation of equivalent actions for a single rule. An heuristic to derive a decision tree for such decision tables was given, without guarantees on how good the derived tree was.…”

Section: B Evaluation Criteriamentioning

confidence: 99%

An Overview of Recent Machine Learning Strategies in Data Mining

Battula¹,

Prasad²

2013

IJACSA

View full text Add to dashboard Cite

Abstract-Most of the existing classification techniquesconcentrate on learning the datasets as a single similar unit, in spite of so many differentiating attributes and complexities involved. However, traditional classification techniques, require to analysis the dataset prior to learning and for not doing so they loss their performance in terms of accuracy and AUC. To this end, many of the machine learning problems can be very easily solved just by careful observing human learning and training nature and then mimic the same in the machine learning. This paper presents an updated literature survey of current and novel machine learning strategies inducing models efficiently for supervised and unsupervised learning in data mining.

show abstract

Optimized Block-Based Connected Components Labeling With Decision Trees

Cited by 155 publications

References 25 publications

A Bathymetry- and Reflectivity-Based Approach for Seafloor Segmentation

A Bathymetry- and Reflectivity-Based Approach for Seafloor Segmentation

Gpu-Based Object Identification in Large-Scale Images for Real-Time Radar Signal Analysis

An Overview of Recent Machine Learning Strategies in Data Mining

Contact Info

Product

Resources

About