Sample stimuli

sample 0 sample 1 sample 2 sample 3 sample 4 sample 5 sample 6 sample 7 sample 8 sample 9

How to use

from brainscore_vision import load_benchmark
benchmark = load_benchmark("MajajHong2015.IT-pls")
score = benchmark(my_model)

Model scores

Min Alignment Max Alignment

Rank

Model

Score

1
.574
2
.572
3
.569
4
.568
5
.566
6
.566
7
.562
8
.561
9
.561
10
.561
11
.560
12
.560
13
.560
14
.559
15
.559
16
.559
17
.558
18
.556
19
.555
20
.555
21
.555
22
.554
23
.554
24
.554
25
.554
26
.554
27
.553
28
.553
29
.553
30
.552
31
.552
32
.552
33
.552
34
.552
35
.552
36
.551
37
.551
38
.550
39
.550
40
.550
41
.549
42
.549
43
.549
44
.549
45
.549
46
.549
47
.548
48
.548
49
.548
50
.547
51
.547
52
.547
53
.547
54
.547
55
.547
56
.546
57
.546
58
.546
59
.546
60
.545
61
.545
62
.545
63
.545
64
.545
65
.544
66
.544
67
.543
68
.543
69
.543
70
.543
71
.543
72
.543
73
.543
74
.542
75
.542
76
.541
77
.541
78
.541
79
.541
80
.541
81
.541
82
.541
83
.541
84
.541
85
.541
86
.541
87
.541
88
.540
89
.540
90
.540
91
.540
92
.539
93
.539
94
.539
95
.539
96
.538
97
.538
98
.536
99
.536
100
.536
101
.536
102
.535
103
.535
104
.535
105
.535
106
.535
107
.535
108
.535
109
.535
110
.534
111
.534
112
.534
113
.534
114
.534
115
.534
116
.533
117
.533
118
.533
119
.533
120
.533
121
.532
122
.532
123
.532
124
.532
125
.532
126
.531
127
.531
128
.531
129
.531
130
.531
131
.531
132
.530
133
.530
134
.530
135
.530
136
.529
137
.529
138
.529
139
.528
140
.528
141
.528
142
.527
143
.527
144
.527
145
.527
146
.527
147
.527
148
.526
149
.526
150
.526
151
.525
152
.525
153
.524
154
.523
155
.523
156
.523
157
.522
158
.522
159
.521
160
.521
161
.520
162
.520
163
.520
164
.520
165
.520
166
.519
167
.519
168
.518
169
.518
170
.518
171
.518
172
.518
173
.517
174
.517
175
.517
176
.517
177
.517
178
.517
179
.516
180
.516
181
.516
182
.515
183
.515
184
.515
185
.515
186
.515
187
.515
188
.514
189
.513
190
.513
191
.513
192
.513
193
.512
194
.512
195
.512
196
.512
197
.512
198
.511
199
.510
200
.510
201
.510
202
.510
203
.509
204
.508
205
.508
206
.508
207
.508
208
.508
209
.508
210
.508
211
.508
212
.508
213
.508
214
.508
215
.508
216
.508
217
.507
218
.507
219
.507
220
.507
221
.506
222
.506
223
.506
224
.506
225
.506
226
.505
227
.505
228
.505
229
.504
230
.504
231
.504
232
.503
233
.503
234
.503
235
.502
236
.502
237
.501
238
.501
239
.501
240
.501
241
.501
242
.499
243
.499
244
.499
245
.499
246
.499
247
.499
248
.498
249
.498
250
.496
251
.496
252
.496
253
.496
254
.495
255
.494
256
.494
257
.494
258
.493
259
.493
260
.493
261
.493
262
.492
263
.490
264
.488
265
.488
266
.487
267
.487
268
.487
269
.486
270
.485
271
.485
272
.485
273
.483
274
.483
275
.483
276
.482
277
.481
278
.480
279
.479
280
.479
281
.477
282
.477
283
.476
284
.475
285
.475
286
.474
287
.473
288
.472
289
.472
290
.472
291
.471
292
.471
293
.470
294
.470
295
.469
296
.469
297
.468
298
.467
299
.467
300
.466
301
.465
302
.464
303
.464
304
.462
305
.461
306
.461
307
.460
308
.459
309
.459
310
.458
311
.457
312
.457
313
.457
314
.456
315
.455
316
.455
317
.455
318
.455
319
.455
320
.455
321
.455
322
.455
323
.455
324
.455
325
.454
326
.451
327
.450
328
.449
329
.447
330
.447
331
.444
332
.443
333
.443
334
.442
335
.441
336
.439
337
.437
338
.437
339
.436
340
.436
341
.435
342
.434
343
.434
344
.432
345
.425
346
.424
347
.423
348
.417
349
.414
350
.414
351
.414
352
.413
353
.411
354
.410
355
.408
356
.408
357
.406
358
.405
359
.404
360
.403
361
.401
362
.401
363
.400
364
.399
365
.395
366
.388
367
.386
368
.382
369
.380
370
.380
371
.371
372
.366
373
.357
374
.351
375
.351
376
.350
377
.344
378
.339
379
.332
380
.325
381
.320
382
.315
383
.307
384
.285
385
.283
386
.278
387
.275
388
.269
389
.263
390
.260
391
.257
392
.257
393
.254
394
.252
395
.251
396
.250
397
.250
398
.246
399
.245
400
.240
401
.215
402
.215
403
.214
404
.212
405
.207
406
.206
407
.206
408
.192
409
.184
410
.174
411
.173
412
.165
413
.159
414
.142
415
.133
416
.132
417
.130
418
.122
419
.109
420
.039
421
.027
422
.026
423
.015
424
.015
425
X
426
X
427
X
428
X
429
X
430
X
431
X
432
X
433
X
434
X
435
X
436
X
437
X
438
X
439
X
440
X
441
X
442
X
443
X
444
X
445
X
446
X
447
X
448
X
449
X
450
X
451
X
452
X
453
X

Benchmark bibtex

@article {Majaj13402,
            author = {Majaj, Najib J. and Hong, Ha and Solomon, Ethan A. and DiCarlo, James J.},
            title = {Simple Learned Weighted Sums of Inferior Temporal Neuronal Firing Rates Accurately Predict Human Core Object Recognition Performance},
            volume = {35},
            number = {39},
            pages = {13402--13418},
            year = {2015},
            doi = {10.1523/JNEUROSCI.5181-14.2015},
            publisher = {Society for Neuroscience},
            abstract = {To go beyond qualitative models of the biological substrate of object recognition, we ask: can a single ventral stream neuronal linking hypothesis quantitatively account for core object recognition performance over a broad range of tasks? We measured human performance in 64 object recognition tests using thousands of challenging images that explore shape similarity and identity preserving object variation. We then used multielectrode arrays to measure neuronal population responses to those same images in visual areas V4 and inferior temporal (IT) cortex of monkeys and simulated V1 population responses. We tested leading candidate linking hypotheses and control hypotheses, each postulating how ventral stream neuronal responses underlie object recognition behavior. Specifically, for each hypothesis, we computed the predicted performance on the 64 tests and compared it with the measured pattern of human performance. All tested hypotheses based on low- and mid-level visually evoked activity (pixels, V1, and V4) were very poor predictors of the human behavioral pattern. However, simple learned weighted sums of distributed average IT firing rates exactly predicted the behavioral pattern. More elaborate linking hypotheses relying on IT trial-by-trial correlational structure, finer IT temporal codes, or ones that strictly respect the known spatial substructures of IT ({	extquotedblleft}face patches{	extquotedblright}) did not improve predictive power. Although these results do not reject those more elaborate hypotheses, they suggest a simple, sufficient quantitative model: each object recognition task is learned from the spatially distributed mean firing rates (100 ms) of \~{}60,000 IT neurons and is executed as a simple weighted sum of those firing rates.SIGNIFICANCE STATEMENT We sought to go beyond qualitative models of visual object recognition and determine whether a single neuronal linking hypothesis can quantitatively account for core object recognition behavior. To achieve this, we designed a database of images for evaluating object recognition performance. We used multielectrode arrays to characterize hundreds of neurons in the visual ventral stream of nonhuman primates and measured the object recognition performance of \>100 human observers. Remarkably, we found that simple learned weighted sums of firing rates of neurons in monkey inferior temporal (IT) cortex accurately predicted human performance. Although previous work led us to expect that IT would outperform V4, we were surprised by the quantitative precision with which simple IT-based linking hypotheses accounted for human behavior.},
            issn = {0270-6474},
            URL = {https://www.jneurosci.org/content/35/39/13402},
            eprint = {https://www.jneurosci.org/content/35/39/13402.full.pdf},
            journal = {Journal of Neuroscience}}

Ceiling

0.82.

Note that scores are relative to this ceiling.

Data: MajajHong2015.IT

2560 stimuli recordings from 168 sites in IT

Metric: pls