initial commit

.gitignore

@@ -0,0 +1,4 @@
+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata

README.md

@@ -0,0 +1,18 @@
+# An OCR Benchmarking Experiment
+
+This repository holds replication materials for the manuscript "OCR with Tesseract, Amazon Textract, and Google Document AI: A Benchmarking Experiment". It contains:
+
+- The .RMD file of the manuscript with R code for all the figures.
+- 51,304 .TXT files with the text output from all the OCR processing requests.
+- A .CSV file with word and character accuracy rates for all the OCR output.
+
+The raw image test materials reside in a separate repository for space reasons.
+
+From the abstract: 
+
+>This article reports a benchmarking experiment comparing the performance of Tesseract, Amazon Textract, and Google Document AI on images of English and Arabic text. English-language book scans (n=322) and Arabic-language article scans (n=100) were replicated 43 times with different types of artificial noise for a corpus of 18,568 documents, generating 51,304 process requests. Document AI delivered the best results, and the server-based processors (Textract and Document AI) were substantially more accurate than Tesseract, especially on noisy documents. Accuracy for English was considerably better than for Arabic.
+
+Core results: 
+\
+\
+![](images/fig4.png)

noisy-ocr-benchmark.Rproj

@@ -0,0 +1,13 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX

ocr_output/old_books/01_col/dai/01_col_a006.tiff.pdf_dai.txt

@@ -0,0 +1,17 @@
+When this book was written, the writer was
+under the supposition then generally current that
+the Armenian Massacres of April, 1909, in Cilicia
+were instigated by Abdul Hamid and his Yildiz
+Clique. Babikian Effendi, the Armenian deputy
+who went to Adana from Constantinople to in-
+vestigate into the massacres, plainly reported that
+all investigations had failed to trace them to Abd-ul
+Hamid and his Yildiz Clique. Babikian Effendi,
+as was to be expected, died suddenly on his return
+to Constantinople, but later on it became known
+that the massacres of April, 1909, had been planned,
+prepared, organized and carried into execution by
+the Constitutional Government of what has been
+called "Liberal Turks" or "Young Turks."
+VE
+

ocr_output/old_books/01_col/dai/01_col_a013.tiff.pdf_dai.txt

@@ -0,0 +1,30 @@
+WHY AND WHEREFORE.
+In making a study of my race, I have found three marked characteristics
+Intelligence Energy-Industry. Combined with these three characteristics
+is an intense Love of Nationality. We live in a complex world. In an
+independent people these characteristics and this sentiment are laudable
+Virtues. In a subject people they are Crimes.
+After I had laid this bitter Truth to heart, I did not have to seek for the
+Why and Wherefore of the Armenian Massacres.
+The Armenian Massacres stand without their parallel in history. The
+human mind staggers to contemplate the fiendish orgies of which they have
+been the victims, and no pen can describe their horrors and this helpless
+christian people are to-day in the same deadly peril as they have been since
+the famous Treaty of Berlin consigned them bound hand and foot to the
+mercy of their executioners.
+The Armenians may be led again "as sheep to the slaughter" and the
+work of extermination may be completed—Jesus Christ was crucified on
+Calvary and the servant is not greater than his Lord-but the work of their
+extermination can only be completed when the evil influences in the Turkish
+Empire have reached their culminating point. Hitherto the Powers of
+Europe have by their jealousies and rivalries cultivated these evil influences,
+they have watered them and made them grow, but when their culminating
+point is reached, they must re-act on Christendom and the natural
+consequence must follow. Those who sow the wind, must reap the whirl-
+wind. It is in the natural order of things.
+1 will allow that Liberty, Justice, Equality, Fraternity are the watchwords
+of Young Turkey, but Young Turkey is only a small minority; the great
+majority of the Turkish nation are not Young Turks.
+The question therefore resolves itself into this critical point: "What will
+Christendom do even now?”
+

ocr_output/old_books/01_col/dai/01_col_a014.tiff.pdf_dai.txt

@@ -0,0 +1,50 @@
+BĪTLIS
+NIGDEH
+EV
+BLACK
+EBIZO N D
+KARA
+BULBURT
+I A
+E
+SIVAS
+PENZERUM
+RUM
+q+
+NEYSHE
+0⁰
+IT
+AKSERA
+ESTAR
+VAN
+IS
+IA'R B EKR
+HADJIM
+MARASH
+MAOURET
+DIARBEKR
+TARSUSADANA
+ASIA
+AYA
+ANTAB
+BIREJIK
+ALEXANDRETTA
+ANTIOCH
+ALEPPO
+MOSU
+MEDITERRANEOSEA
+ABUD SAIDEH
+SCENE OF THE MASSACRES IN ASIA MINOR.
+The trouble began in Adana. An armed mob strengthened and augmented by soldiers
+fell in overwhelming numbers upon the unarmed Christians. The Armenian population of
+Antioch and vicinity were practically wiped out and the Armenian villages in the
+Alexandretta district destroyed with immense loss of life. Hadjim, Kessab and the neigh-
+bouring villages were burned. The Armenian quarter in Tarsus was ruined and ill-omened
+Marash stained again with the blood of thousands of Armenians. Zeitoon was desolated. The
+entire population of Kirikon between Aleppo and Alexandretta were massacred to the last
+babe. The mob and the soldiers burned what they could not carry away, so that the material
+loss has been enormous. In place of the former abundance and thriving industries there are
+instead desolated provinces and the charred and blackened remains of pillaged and ruined
+homes, and the residue of those who escaped massacre are reduced to homelessness and
+starvation,
+

ocr_output/old_books/01_col/dai/01_col_a015.tiff.pdf_dai.txt

@@ -0,0 +1,30 @@
+DISINTERESTED EVIDENCE.
+66
+I have thought it advisable to insert a few extracts from accounts of
+the Massacres of April, 1909, given by disinterested witnesses.
+We are having a perfectly hideous time here. Thousands have been murdered-25,000
+in this province they say; but the number is probably greater, for every Christian village
+was wiped out In Adana about 5000 have perished. After Turks and Armenians had
+made peace, the Turks came in the night with hose and kerosene, and set fire to what remained
+of the Armenian quarter. Next day the French and Armenian schools were fired. Nearly every-
+one in the Armenian school perished, anybody trying to escape being shot down by thesoldiers."
+"The Turkish Authorities do nothing except arrest unoffending Armenians, from whom
+by torture they extort the most fanciful confessions. Even the wounded are not safe from
+their injustice. A man was being carried in to me yesterday when he was seized and taken
+off to gaol. I dare not think what his fate may be.
+"For fiends incarnate commend me to the Turks. Nobody is safe from them. They
+murder babies in front of their mothers; they half murder men, and violate the wives while
+the husbands are lying there dying in pools of blood.”
+"The authorities did nothing, and the soldiers were worse than the crowd, for they were
+better armed. One house in our quarter was burned with 115 people inside. We counted
+the bodies. The soldiers set fire to the door, and as the windows had iron bars, nobody
+could get out. Everybody in the house was roasted alive. They were all women and
+children and old people.”—Extract from letter of Mrs. Doughty-Wylie, wife of British Consul
+at Adana ; published in the London “Daily Mail.”
+961
+IN THIS HOUSE 115 WOMEN AND CHILDREN WERE ROASTED ALIVE.
+History repeats itself. In 1895 Turkish soldiers fell upon seventy to eighty young
+women and girls in a church, where they had fled for refuge, and after hideously outraging
+them, barricaded them in, setting fire to the building at the same time, and derisively shouting
+to their victims as they were being roasted alive, to call upon their Christ to save them now
+

ocr_output/old_books/01_col/dai/01_col_a017.tiff.pdf_dai.txt

@@ -0,0 +1,37 @@
+of blackened ruins, the stark walls of the churches and houses rising up out of the ashes and
+charred timbers heaped on every side. What must it mean to the five thousand men and
+women and little children who have survived a painful flight to the seacoast and have now
+returned to their mountain home, only to find their houses sacked and burned! There were
+nine Christian villages which elustered about Kessab in the valleys below. Several of these
+have been completely destroyed by fire. All have been plundered and the helpless people
+driven out or slain.”
+“Can you imagine the feelings of the Kessab people as they climbed on foot the long trail
+up the mountain, and then as they came over the ridge into full view of their charred and
+ruined dwellings? Their stores of wheat, barley and rice had been burned ; clothing, cooking
+utensils, furniture and tools had gone; their goats, cows and mules had been stolen; their silk
+industries stamped out; their beloved churches reduced to smouldering heaps. The bodies of
+their friends and relatives who had been killed had not been buried. And yet the love of
+home is so strong that the people have settled down there with the determination to clear up
+the debris and rebuild their houses.”—Extracts from "The Sack of Kessab," Stephen Van R.
+Trowbridge.
+As these sheets are going through the press there comes news of famine
+at Zeitoon. The Rev. F. W. Macullum, American Missionary at Marash,
+writes to the Rev. W. W. Peet, American Missionary at Constantinople, that
+12,000 souls in and around Zeitoon are dying of hunger; they are wandering
+about in rags, mixing bran and water, and cooking and eating it, if they can
+get even that. Rev. Macullum adds, “The same story comes to us from all
+sides. As we foresaw all along, from now on the distress will be greatest.
+If 50,000 were massacred, the list of those who have died and are dying
+of homelessness and starvation will exceed 150,000. It is true; and the
+numbers are not exaggerated. Last year the people reaped no harvest, and
+this year there are no sowings.
+The latest news is that Mush, a prosperous Armenian village that had
+escaped the desolation of the massacres, has been plundered in a night attack
+by armed Kurds, and the villagers are now reduced to extreme distress. Before
+the outbreak the Armenian patriarchal vicar at Mush had repeatedly appealed
+to the Armenian Patriarch at Constantinople, and the Armenian Patriarch
+had repeatedly appealed to the Authorities at Constantinople asking protec-
+tion for the villagers of Mush as a Kurdish attack was apprehended. It is
+evident that the authorities at Constantinople are unable to protect thriving
+Armenian villages from Kurdish and Turkish raiders.
+

ocr_output/old_books/01_col/dai/01_col_a018.tiff.pdf_dai.txt

@@ -0,0 +1,10 @@
+PREFACE TO 2ND PRINTING.
+The first and second parts of this little book were written and printed
+in pamphlet form for circulation in the United States, shortly after the
+Adana Massacres of April, 1909. I have now thought it advisable to add
+a Supplement of a short history of the Origin of the Armenians and the
+Introduction and Revival of Christianity in Armenia.
+The illustrations and the extracts from the periodicals “Harper's
+Monthly," "The Wide World" and the "Cosmopolitan” have been added
+to the 2nd printing.
+

ocr_output/old_books/01_col/dai/01_col_a019.tiff.pdf_dai.txt

@@ -0,0 +1,36 @@
+INTRODUCTION
+TO 2ND PRINTING.
+My object in writing this little book is to lay the hard case of my un-
+fortunate race before the men and women of the United States; since it is
+from the United States that the American Missionaries have gone forth, who
+have been the only helping influence from without for my suffering people in
+Asiatic Turkey. To the earnest and devoted men and women of the
+American Missions, we Armenians owe a debt of gratitude which we can
+never repay.
+If in the contents of the pages of this little book I have exaggerated
+Facts by one whit or one iota, if I have deviated by one hair's breadth from
+the Truth, I stand to be judged.
+"God save us from another Adana, but the sword of Islam has not been
+dulled" was one of the clarion notes sounded at the Sixth International
+Convention of the Student Volunteer Movement, which was held at Rochester,
+New York. The man who sounded that clarion note knew Islam, and
+because knowing of my own knowledge that the sword of Islam has not been
+dulled, I tremble lest its sharp edge fall once more on the neck of my help-
+less race.
+If I knew and felt sure in mine own heart that the sword of Islam
+was dulled, I would be content to let bygones be bygones, and to hold my
+peace and be silent for ever.
+Like the sudden explosion of a volcano in the physical world, comes the
+explosion of a Turkish Massacre of Armenians in the moral world. It comes
+just in that way; the subterranean fires are always there, but all of a sudden
+the sulphur flames of religious fanaticism burst, the lava floods of race hatred
+and lust of plunder, break forth and run in fiery streams; the unfortunate
+victims are pounced upon, swooped upon, pillaged, plundered, butchered,
+slaughtered, subjected to outrages so hideous, cruel, loathsome, and revolt-
+ing, that no pen could depict their horrible realities and the details can never
+go into print. The human mind is staggered and asks itself the question if
+even the imaginations of fiends and devils could originate such horrors.
+Then this orgy of the human fiends is arrested. For the time being the
+appetite for blood, lust, and plunder is satisfied; for the time being, the eye
+is content with the scenes of havoc and desolation lying under the sun; the
+

ocr_output/old_books/01_col/dai/01_col_a020.tiff.pdf_dai.txt

@@ -0,0 +1,41 @@
+10
+smell of corpses is in the air, the odor from the carcases of the “christian
+swine” reek in the nostrils of the Turk, he turns away, his jaws dripping with
+blood, and rests to couch for a future spring. We have seen that sort of an
+end to the tragedy of a tiger's victim: the tiger has eaten his fill, he rests,
+to keep guard over the crunched bones and mangled bits of bloody flesh that
+bestrew the earth. So also now there is a residue left of those that have
+served as the meat and wine of this devil's feast; the demons have gorged
+themselves over the banquet, and now there are left over the broken remains
+of the banquet, the miserable residue homeless and destitute.
+Civilized nations have received a temporary moral shock, like a shock
+that spreads from the centre of an explosion; the electric vibration running
+far and wide from the scene of the centre of devastation. There are among
+these civilized nations generous and kind-hearted people who open their
+purse strings; they give money to purchase shelter, food and clothing for
+these homeless, naked and, hungry beggars, made homeless, naked and
+hungry through no fault of their own. But oh! ye generous and kind heart-
+ed people can any power under heaven assuage the heart anguish of this
+miserable residue? Can they be made by any means of human comfort to
+forget the black horrors or recover from the effects of the fires of the hideous
+affliction through which they have passed? What is there left for a woman
+who has seen with her own eyes the slaughter and heard with her own ears
+the dying cry of her murdered child? even her reason must give way under
+the stress of her anguish. All ye who are mothers, I appeal to you, for one
+moment to put yourselves in the place of thousands of such mothers, in
+whose hearts the same mother's love burns as in yours, and then measure
+the depth of their agony.
+Generous and kind hearted people who open your purse strings; would
+to God I entreat, ye would raise up your voices and demand that this hideous
+slaughter and oppression of a helpless christian race should cease. Would
+to God I entreat, ye would raise up your voices and demand that this people
+of an industrious, intelligent christian race, robust in mind and body, should
+be let to live. Would to God I entreat, that ye would raise up your voices
+and demand for them that security of life and property to which they are
+entitled just as equally as all other peoples.
+Public Sentiment has done great things in the world's history. Public
+Sentiment liberated Greece, The Lebanon, The Balkan States from Turkish
+Oppression. Slavery was abolished in the United States through Public
+Sentiment but alas! does Public Sentiment sleep for this helpless Christian
+race. Are they not God's creatures? have they not a right to live on God's
+

ocr_output/old_books/01_col/dai/01_col_a021.tiff.pdf_dai.txt

@@ -0,0 +1,42 @@
+11
+earth as other nations? Does Humanity,' does Christianity allow that tender
+babes and children should be hideously and horribly mutilated and butchered
+before the eyes of their mothers, or that the ears of mothers should be rent
+with the cries of the dying agony of their murdered children ? Does
+Humanity, does Christianity allow that helpless women should be forcibly
+subjected to the most hideous, the most loathsome, the most revolting, and
+the most cruel outrages? Does Humanity, does Christianity, allow all this?
+Christian Governments have organized a Hague Conference of Peace
+and Civilization, but they have. closed its doors to the cause of a bleeding
+christian race groaning under the yoke of the cruellest oppressors that the
+world has yet known. Christian men and women have held up their hands
+in horror at the Indian Juggernauth; but alas! the political wheels of
+Christian Governments have been a Greater Juggernauth for a helpless
+christian race. It is by Christian Governments that “we are made as the
+filth of the world, and as the offscouring of all things unto this day.” It is
+as if the answer to our groanings had been made by Christian Governments
+in just these words :
+“We know that you have had frightful grievances, such as have been
+beyond the measure of human endurance. We know that since the Treaty
+of Berlin your history has been written in blood and tears, as the history of
+no other nation has been written before or now. We know that your women
+are subjected to the most revolting and hideous agonies, and your babes
+and children hounded to hideous deaths. We know that the sum total of
+your wrongs and sufferings is so great, that the cry of its anguish is piercing
+the very heavens, but really, our political and commercial jealousies prevent;
+and we each one of us being on the look out lest our separate political and
+commercial interests in the Empire of your oppressors be endangered, can-
+not regard you. It may be the deadliest scandal of Christendom that we
+Christian Powers should be all gathered together, one against another, in the
+Empire of your Oppressors, as eagles gather together round a carcase; but
+really there is no help for it; and if you must die hideously by a hellish
+extermination, why then you must die, and we have to condone your
+hellish
+extermination, for in any case, each one of us must secure his own political
+and commercial interests in this same Empire of your Oppressors.'
+In "Transcaucasia and Ararat," published by Mr. James Bryce in 1876,
+there occurs in the chapter entitled “Some Political Reflections” the follow-
+ing passage :
+"The attention of the West was so much drawn towards Herzegovina
+and Bulgaria by the events of 1876 that the miseries of the Asiatic subjects
+