AudreyBeard
commented
7 years ago Original note:
It seems that chip_uid is set to the exact same (default, '_CHIP(sz750)') for each chip. My guess is that this issue happens at the point of populating the chip table. it seems that 'sz' is short for 'size' and that '750' is the default value for the square root of the area.
Response:
After exploring this further, it seems that if we add a chip the old-fashioned way (using Actions->Add Chip), the chip_uid has the same default name. Because of this, I don't think this is the issue.
probably Deprecated Work:
Following chip_compute2.py
chip_cfg = hs.prefs.chip_cfg
chip_uid = chip_cfg.get_uid()
Leads to HotSpotterAPI.py
hs.prefs.chip_cfg = Config.default_chip_cfg()
To Config.py
def default_chip_cfg(**kwargs):
chip_cfg = ChipConfig(**kwargs)
return chip_cfgAND
class ChipConfig(ConfigBase):
def __init__(cc_cfg, **kwargs):
super(ChipConfig, cc_cfg).__init__(name='chip_cfg')
cc_cfg.chip_sqrt_area = 750
cc_cfg.grabcut = False
cc_cfg.histeq = False
cc_cfg.adapteq = False
cc_cfg.region_norm = False
cc_cfg.rank_eq = False
cc_cfg.local_eq = False
cc_cfg.maxcontrast = False
cc_cfg.update(**kwargs)AND
ConfigBase = Pref
AND
from hscom.Preferences import Pref
mattdioso
Affected Branches:
All Branches Affected
To Reproduce:
Output:
==================== [hs] query database
[mc3] FEAT_UID is different. Need to reload features [mc3] Old: None [mc3] New: _FEAT(hesaff+sift,0_9001)_CHIP(sz750) [hs] unload_cxdata(cx='all') [hs] Unloading all data [hs] finished unloading all data
============================= [cc2] Precomputing chips and loading chip paths: 'test0'
[cc2] chip_uid = '_CHIP(sz750)' [parallel] Already computed 1 compute_chip tasks
tic('Distributing 8 compute_chip: tasks to 8 processes') compute_chip: compute_chip: 8/8 ...toc('Distributing 8 compute_chip: tasks to 8 processes')=0.2567s [xcb] Unknown sequence number while processing queue [xcb] Most likely this is a multi-threaded client and XInitThreads has not been called [xcb] Aborting, sorry about that. python: ../../src/xcb_io.c:274: poll_for_event: Assertion `!xcb_xlib_threads_sequence_lost' failed. [cc2] Done Precomputing chips and loading chip paths [cc2]=============================
============================= [fc2] Precomputing and loading features: 'test0'
[fc2] feat_uid = '_FEAT(hesaff+sift,0_9001)_CHIP(sz750)' [io] fname='kpts_list_FEAT(hesaff+sift,0_9001)_CHIP(sz750)_cids((9,)5l$qyq1{).npy' does not exist [io] fname='desc_list_FEAT(hesaff+sift,0_9001)_CHIP(sz750)_cids((9,)5l$qyq1{).npy' does not exist [fc2] Loading _FEAT(hesaff+sift,0_9001)_CHIP(sz750) individually [parallel] Already computed 0 precompute_hesaff tasks
tic('Distributing 9 precompute_hesaff: tasks to 8 processes') precompute_hesaff: Error in `python': malloc(): smallbin double linked list corrupted: 0x0000000005057510 ======= Backtrace: =========
Error in `python': free(): corrupted unsorted chunks: 0x0000000005057dd0