Inventor · Beijing, CN

Frank Kao-Ping Soong

35Patents
13h-index
36Co-inventors
77Inventor score

Filing activity: Jan 21, 1994 → Feb 17, 2012

Most-cited inventions

PatentTitleAreaCited byStatus
US7689421B2 Voice persona service for embedding text-to-speech features into software programs Physics 297 Active
US7983478B2 Hidden markov model based handwriting/calligraphy generation Physics 284 Active
US7689408B2 Identifying language of origin for words using estimates of normalized appearance frequency Physics 278 Active
US5377301A Technique for modifying reference vector quantized speech feature signals Electricity 240 Expired
US5675704A Speaker verification with cohort normalized scoring Electricity 186 Expired
US8224652B2 Speech and text driven HMM-based body animation synthesis Physics 88 Active
US6009390A Technique for selective use of Gaussian kernels and mixture component weights of tied-mixture hidden Markov models for speech recognition Physics 73 Expired
US8077975B2 Handwriting symbol recognition accuracy using speech input Physics 61 Active
US6138095A Speech recognition Physics 45 Expired
US5680506A Apparatus and method for speech signal analysis Physics 19 Expired
US8244534B2 HMM-based bilingual (Mandarin-English) TTS techniques Physics 19 Active
US9613450B2 Photo-realistic synthesis of three dimensional animation with facial features synchronized with speech Physics 15 Active
US7860716B2 Speech model refinement with transcription error detection Physics 14 Active
US7890325B2 Subword unit posterior probability for measuring confidence Physics 11 Active
US5995926A Technique for effectively recognizing sequence of digits in voice dialing Electricity 10 Expired
US7805004B2 Radical set determination for HMM based east asian character recognition Physics 7 Active
US7848917B2 Common word graph based multimodal input Physics 7 Active
US7844457B2 Unsupervised labeling of sentence level accent Physics 7 Active
US7496513B2 Combined input processing for a computing device Physics 5 Active
US8594993B2 Frame mapping approach for cross-lingual voice transformation Physics 4 Active
US6701291B2 Automatic speech recognition with psychoacoustically-based feature extraction, using easily-tunable single-shape filters along logarithmic-frequency axis Physics 4 Expired
US7977562B2 Synthesized singing voice waveform generator Physics 4 Active
US8340965B2 Rich context modeling for text-to-speech engines Physics 3 Active
US8355917B2 Position-dependent phonetic models for reliable pronunciation identification Physics 3 Active
US7805301B2 Covariance estimation for pattern recognition Physics 2 Active

Source: USPTO / EPO open patent data. Inventor disambiguation is heuristic; counts are objective bibliographic measures.