Matches in SemOpenAlex for { <https://semopenalex.org/work/W2282541982> ?p ?o ?g. }
Showing items 1 to 60 of
60
with 100 items per page.
- W2282541982 abstract "Pattern recognition is the assignment of some sort of label to a given input valueor instance, according to some specific learning algorithm. The recognitionperformance is directly linked with the quality and size of the training data.However, in many real pattern recognition implementations, it is difficult or not soconvenient to collect as many samples as possible for training up the classifier,such as face recognition or Chinese character recognition.In view of the shortage of training samples, the main object of our research is toinvestigate the generation and use of artificial samples for improving therecognition performance. Besides enhancing the learning, artificial samples arealso used in a novel way such that a conventional Chinese character recognizercan read half or combined Chinese character segments. It greatly simplifies thesegmentation procedure as well as reduces the error introduced by segmentation.Two novel generation models have been developed to evaluate the effectivenessof supplementing artificial samples in the training. One model generates artificialfaces with various facial expressions or lighting conditions by morphing andwarping two given sample faces. We tested our face generation model in threepopular 2D face databases, which contain both gray scale and color images.Experiments show the generated faces look quite natural and they improve therecognition rates by a large margin.The other model uses stroke and radical information to build new Chinesecharacters. Artificial Chinese characters are produced by Bezier curves passingthrough some specified points. This model is more flexible in generating artificialhandwritten characters than merely distorting the genuine real samples, with bothstroke level and radical level variations. Another feature of this charactergeneration model is that it does not require any real handwritten character sampleat hand. In other words, we can train the conventional character classifier andperform character recognition tasks without collecting handwritten samples.Experiment results have validated its possibility and the recognition rate is stillacceptable.Besides tackling the small sample size problem in face recognition and isolatedcharacter recognition, we improve the performance of bank check legal amountrecognizer by proposing character segments recognition and applying HiddenMarkov Model (HMM).It is hoped that this thesis can provide some insights for future researches inartificial sample generation, face morphing, Chinese character segmentation andtext recognition or some other related issues." @default.
- W2282541982 created "2016-06-24" @default.
- W2282541982 creator A5069564171 @default.
- W2282541982 date "2015-05-12" @default.
- W2282541982 modified "2023-09-27" @default.
- W2282541982 title "Artificial training samples for the improvement of pattern recognition systems" @default.
- W2282541982 doi "https://doi.org/10.5353/th_b4784964" @default.
- W2282541982 hasPublicationYear "2015" @default.
- W2282541982 type Work @default.
- W2282541982 sameAs 2282541982 @default.
- W2282541982 citedByCount "0" @default.
- W2282541982 crossrefType "dissertation" @default.
- W2282541982 hasAuthorship W2282541982A5069564171 @default.
- W2282541982 hasConcept C153180895 @default.
- W2282541982 hasConcept C154945302 @default.
- W2282541982 hasConcept C31510193 @default.
- W2282541982 hasConcept C41008148 @default.
- W2282541982 hasConcept C50637493 @default.
- W2282541982 hasConcept C50644808 @default.
- W2282541982 hasConcept C77088390 @default.
- W2282541982 hasConcept C88548561 @default.
- W2282541982 hasConcept C89600930 @default.
- W2282541982 hasConcept C95623464 @default.
- W2282541982 hasConceptScore W2282541982C153180895 @default.
- W2282541982 hasConceptScore W2282541982C154945302 @default.
- W2282541982 hasConceptScore W2282541982C31510193 @default.
- W2282541982 hasConceptScore W2282541982C41008148 @default.
- W2282541982 hasConceptScore W2282541982C50637493 @default.
- W2282541982 hasConceptScore W2282541982C50644808 @default.
- W2282541982 hasConceptScore W2282541982C77088390 @default.
- W2282541982 hasConceptScore W2282541982C88548561 @default.
- W2282541982 hasConceptScore W2282541982C89600930 @default.
- W2282541982 hasConceptScore W2282541982C95623464 @default.
- W2282541982 hasLocation W22825419821 @default.
- W2282541982 hasOpenAccess W2282541982 @default.
- W2282541982 hasPrimaryLocation W22825419821 @default.
- W2282541982 hasRelatedWork W142286180 @default.
- W2282541982 hasRelatedWork W1519545881 @default.
- W2282541982 hasRelatedWork W1971442148 @default.
- W2282541982 hasRelatedWork W1999797110 @default.
- W2282541982 hasRelatedWork W2016362806 @default.
- W2282541982 hasRelatedWork W2050182871 @default.
- W2282541982 hasRelatedWork W2095723660 @default.
- W2282541982 hasRelatedWork W2130677977 @default.
- W2282541982 hasRelatedWork W2147600805 @default.
- W2282541982 hasRelatedWork W2159255787 @default.
- W2282541982 hasRelatedWork W2162265202 @default.
- W2282541982 hasRelatedWork W2188101492 @default.
- W2282541982 hasRelatedWork W2791314317 @default.
- W2282541982 hasRelatedWork W3119350346 @default.
- W2282541982 hasRelatedWork W3188260853 @default.
- W2282541982 hasRelatedWork W3192270299 @default.
- W2282541982 hasRelatedWork W1894382562 @default.
- W2282541982 hasRelatedWork W192530680 @default.
- W2282541982 hasRelatedWork W2288869608 @default.
- W2282541982 hasRelatedWork W2398918985 @default.
- W2282541982 isParatext "false" @default.
- W2282541982 isRetracted "false" @default.
- W2282541982 magId "2282541982" @default.
- W2282541982 workType "dissertation" @default.