​​Optical character recognition (Optical Character Recognition (OCR)) has many applications, such as digitizing historical documents, automating processes, and helping visually impaired people read. However, extracting text from images into a digital format is not an easy problem to solve, and the outputs from the OCR frameworks often include errors. The complexity comes from the many variations in (digital) fonts, handwriting, lighting, etc. To tackle this problem, this thesis investigates two different methods for correcting the errors in OCR output. The used dataset consists of Swedish addresses. The methods are therefore applied to postal automation to investigate the usage of these methods for further automating postal work by automatically reading addresses on parcels using OCR. The main method, the lexical implementation, uses a dataset of Swedish addresses so that any valid address should be in this dataset (hence there is a known and limited vocabulary), and misspelled addresses are corrected to the address in the lexicon with the smallest Levenshtein distance. The second approach is to use the same dataset, but with artificial errors, or artificial noise, added. The addresses with this artificial noise are then used together with their correct spelling to train a machine learning model based on Neural machine translation (Neural Machine Translation (NMT)) to automatically correct errors in OCR read addresses. The results from this study could contribute by defining in what direction future work connected to OCR and postal addresses should go. The results were that the lexical implementation outperformed the NMT model. However, more experiments including real data would be required to draw definitive conclusions as to how the methods would work in real-life applications. / Optisk teckenigenkänning (Optical Character Recognition (OCR)) har många användningsområden, till exempel att digitalisera historiska dokument, automatisera processer och hjälpa synskadade att läsa. Att extrahera text från bilder till ett digitalt format är dock inte ett lätt problem att lösa, och utdata från OCR-ramverken innehåller ofta fel. Komplexiteten kommer från de många variationerna i (digitala) typsnitt, handstil, belysning, etc. För att lösa problemet undersöker den här avhandling två olika metoder för att rätta fel i OCR-utdata. Det använda datasetet består av svenska adresser. Metoderna tillämpas därför på postautomatisering för att undersöka användningen av dessa metoder för att ytterligare automatisera postarbetet genom att automatiskt läsa adresser på paket med OCR. Den första metoden, den lexikaliska metoden, använder en datauppsättning av svenska adresser så att alla giltiga adresser bör finnas i denna datauppsättning (därav finns det ett känt och begränsat ordförråd). Denna datauppsättning används sedan som en ordbok för att hitta adressen med det minsta Levenshtein-avståndet till någon felstavad adress. Det andra tillvägagångssättet använder samma datauppsättning, men med artificiella fel tillagda. Adresserna med dessa artificiella fel används sedan tillsammans med deras korrekta stavning för att träna en Neural Machine Translation (NMT)-modell för att automatiskt korrigera fel i OCR-lästa adresser. Resultaten från denna studie skulle kunna bidra genom att definiera i vilken riktning framtida arbete kopplat till OCR och postadresser ska gå. Resultaten var att den lexikaliska metoden presterade bättre än NMT-modellen. Fler experiment gjorde med verklig data skulle dock behövas för att dra definitiva slutsatser om hur metoderna skulle fungera i verkliga tillämpningar.

長久以來，序列式詞彙處理模式以及互動式詞彙處理模式經常爭論孰為合理的語誤生成模型。這份研究企圖利用心理學實證角度探究此議題。原屬於心理學範疇的史楚普技術(1935)或可提供語言學多方面的視角，尤其在我們處理顏色詞彙時詞彙譯碼歷程上與視覺表徵之間的關係。因此我們試圖將此經典心理學技術轉化成宜於語言學研究的語誤誘發實驗。藉由控制變項間的語音近似度，創制了四項實驗，分別是：顏色唸名實驗、顏色誦讀實驗、史楚普唸名實驗、同音詞唸名實驗。結果顯示，即便受試者對語音近似度並未在單項實驗中產生語誤數量與反應時間的差異，但它在不同實驗任務之間卻誘使受試者產生不一樣的反應結果，包括數量上及時間上的顯著差異。 此外，我們針對語誤，分析其與目標字之間的關係，研判部分音韻效應是否對詞彙處理網絡造成顯著影響。結果顯示，史楚普效應、音韻結構、音法限制以及聲調效應在語誤數量中均出現顯著的生成量；音節首、韻部以及母音則並未出現顯著效應。史楚普技術除了在語誤證據中為文字閱讀提供自動化效應的解釋外，也讓我們看到音韻效應在不同視覺任務指派中產生顯著差異的結果。如此看來，互動式模型提供了較為簡潔的解釋。來自二元視覺刺激(視覺色彩與顏色詞)的詞彙競爭，該框架可以合理提供理論基礎，並解釋不同視覺任務之間音韻關係依存度的不同。 另一方面，我們亦援引同樣技術設計出以音韻單位為導向的唸名實驗，討論語言輸出前規劃單位之議題。我們發現，僅含聲調的音節以及無聲調的音節可在詞彙網絡中做為語言規劃之心理處理單位，促進詞彙處理效能。其他單位如音節首、介音、母音、韻尾、聲調則無法出現顯著效應。同時亦於實驗中發現，中文聲調於規劃階段時，應是屬於詞彙結構上的聲調，而非純粹的聲韻調。 / Serial-ordering model and interactive processing model have long been discussed as whether people process languages in a sequential level of processing or the consequence of levels interacting altogether by looking at speech errors. Stroop technique (1935) in psychology could give linguistics some insights on the relation between lexical encoding and visual representation when people process colors. We tried to adapt this classical psychological experiment for an experiment of speech error elicitation. By means of controlling the phonological similarity, we created four experiments: color naming, color reading, Stroop naming, and homophonous naming tasks. The result first showed that even though phonological similarity did not induce significant difference in error amount and response time within single task, it still caused significant difference in these results among these tasks. Second, after analyzing the linguistic relation between targets and speech errors, we found that Stroop effect, syllable structure, phonotactic regularity, and tone induced apparent effects in error generation, while initial, rhyme, and vowel did not. Stroop effect not only provided the evidence from speech errors where character representation was an autonomous mechanism in lexical process, but also provided a fact that phonological effect would impacted differently on amount of error according to the type of visual task. It seems that interactive account could help explain the result easily, as the competition from dual visual inputs could be given a theoretical basis to account for the phonological dependency according to certain visual task which subjects were assigned to. Finally, we also discussed the issue of advanced planning unit in lexical process. In the shared unit task with Stroop technique, we found that only the units of tonal syllable and bare syllable could serve as possible planning units in lexical network, and tone in Chinese should be attributed to a type of lexical tone in planning, rather than a pure phonological tone. To sum up, the purpose of the study attempts to provide empirical evidence to examine the above issues.