Mercurial > hgrepos > Python2 > PyMuPDF

comparison mupdf-source/thirdparty/tesseract/src/classify/cluster.h @ 2:b50eed0cc0ef upstream

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4. The directory name has changed: no version number in the expanded directory now.
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:43:07 +0200
parents
children
comparison
equal deleted inserted replaced
1:1d09e1dec1d9 2:b50eed0cc0ef
1 /******************************************************************************
2 ** Filename: cluster.h
3 ** Purpose: Definition of feature space clustering routines
4 ** Author: Dan Johnson
5 **
6 ** (c) Copyright Hewlett-Packard Company, 1988.
7 ** Licensed under the Apache License, Version 2.0 (the "License");
8 ** you may not use this file except in compliance with the License.
9 ** You may obtain a copy of the License at
10 ** http://www.apache.org/licenses/LICENSE-2.0
11 ** Unless required by applicable law or agreed to in writing, software
12 ** distributed under the License is distributed on an "AS IS" BASIS,
13 ** WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 ** See the License for the specific language governing permissions and
15 ** limitations under the License.
16 *****************************************************************************/
17
18 #ifndef CLUSTER_H
19 #define CLUSTER_H
20
21 #include "kdtree.h"
22 #include "oldlist.h"
23
24 namespace tesseract {
25
26 struct BUCKETS;
27
28 #define MINBUCKETS 5
29 #define MAXBUCKETS 39
30
31 /*----------------------------------------------------------------------
32 Types
33 ----------------------------------------------------------------------*/
34 struct CLUSTER {
35 CLUSTER(size_t n) : Mean(n) {
36 }
37
38 ~CLUSTER() {
39 delete Left;
40 delete Right;
41 }
42
43 bool Clustered : 1; // true if included in a higher cluster
44 bool Prototype : 1; // true if cluster represented by a proto
45 unsigned SampleCount : 30; // number of samples in this cluster
46 CLUSTER *Left; // ptr to left sub-cluster
47 CLUSTER *Right; // ptr to right sub-cluster
48 int32_t CharID; // identifier of char sample came from
49 std::vector<float> Mean; // mean of cluster - SampleSize floats
50 };
51 using SAMPLE = CLUSTER; // can refer to as either sample or cluster
52
53 typedef enum { spherical, elliptical, mixed, automatic } PROTOSTYLE;
54
55 struct CLUSTERCONFIG { // parameters to control clustering
56 PROTOSTYLE ProtoStyle; // specifies types of protos to be made
57 float MinSamples; // min # of samples per proto - % of total
58 float MaxIllegal; // max percentage of samples in a cluster which
59 // have more than 1 feature in that cluster
60 float Independence; // desired independence between dimensions
61 double Confidence; // desired confidence in prototypes created
62 int MagicSamples; // Ideal number of samples in a cluster.
63 };
64
65 typedef enum { normal, uniform, D_random, DISTRIBUTION_COUNT } DISTRIBUTION;
66
67 union FLOATUNION {
68 float Spherical;
69 float *Elliptical;
70 };
71
72 struct PROTOTYPE {
73 bool Significant : 1; // true if prototype is significant
74 bool Merged : 1; // Merged after clustering so do not output
75 // but kept for display purposes. If it has no
76 // samples then it was actually merged.
77 // Otherwise it matched an already significant
78 // cluster.
79 unsigned Style : 2; // spherical, elliptical, or mixed
80 unsigned NumSamples : 28; // number of samples in the cluster
81 CLUSTER *Cluster; // ptr to cluster which made prototype
82 std::vector<DISTRIBUTION> Distrib; // different distribution for each dimension
83 std::vector<float> Mean; // prototype mean
84 float TotalMagnitude; // total magnitude over all dimensions
85 float LogMagnitude; // log base e of TotalMagnitude
86 FLOATUNION Variance; // prototype variance
87 FLOATUNION Magnitude; // magnitude of density function
88 FLOATUNION Weight; // weight of density function
89 };
90
91 struct CLUSTERER {
92 int16_t SampleSize; // number of parameters per sample
93 PARAM_DESC *ParamDesc; // description of each parameter
94 int32_t NumberOfSamples; // total number of samples being clustered
95 KDTREE *KDTree; // for optimal nearest neighbor searching
96 CLUSTER *Root; // ptr to root cluster of cluster tree
97 LIST ProtoList; // list of prototypes
98 uint32_t NumChar; // # of characters represented by samples
99 // cache of reusable histograms by distribution type and number of buckets.
100 BUCKETS *bucket_cache[DISTRIBUTION_COUNT][MAXBUCKETS + 1 - MINBUCKETS];
101 };
102
103 struct SAMPLELIST {
104 int32_t NumSamples; // number of samples in list
105 int32_t MaxNumSamples; // maximum size of list
106 SAMPLE *Sample[1]; // array of ptrs to sample data structures
107 };
108
109 // low level cluster tree analysis routines.
110 #define InitSampleSearch(S, C) (((C) == nullptr) ? (S = NIL_LIST) : (S = push(NIL_LIST, (C))))
111
112 /*--------------------------------------------------------------------------
113 Public Function Prototypes
114 --------------------------------------------------------------------------*/
115 TESS_API
116 CLUSTERER *MakeClusterer(int16_t SampleSize, const PARAM_DESC ParamDesc[]);
117
118 TESS_API
119 SAMPLE *MakeSample(CLUSTERER *Clusterer, const float *Feature, uint32_t CharID);
120
121 TESS_API
122 LIST ClusterSamples(CLUSTERER *Clusterer, CLUSTERCONFIG *Config);
123
124 TESS_API
125 void FreeClusterer(CLUSTERER *Clusterer);
126
127 TESS_API
128 void FreeProtoList(LIST *ProtoList);
129
130 void FreePrototype(void *arg); // PROTOTYPE *Prototype);
131
132 CLUSTER *NextSample(LIST *SearchState);
133
134 float Mean(PROTOTYPE *Proto, uint16_t Dimension);
135
136 float StandardDeviation(PROTOTYPE *Proto, uint16_t Dimension);
137
138 TESS_API
139 int32_t MergeClusters(int16_t N, PARAM_DESC ParamDesc[], int32_t n1, int32_t n2, float m[],
140 float m1[], float m2[]);
141
142 } // namespace tesseract
143
144 #endif